JP6519682B2 - Measuring device, measuring method and measuring program - Google Patents

Description

  The invention disclosed herein relates to a measurement device, a measurement method, and a measurement program.

  2. Description of the Related Art In recent years, various sensors have been mounted on information processing apparatuses such as devices that measure the movement of moving objects, as their functions become more sophisticated. For example, a device having an acceleration sensor or an angular velocity sensor determines the current state of the device based on the output value of the sensor. And the method etc. of determining the walking state and posture of the person who carries an information processing apparatus are proposed using the output value of an acceleration sensor or an angular velocity sensor (refer to patent documents 1 and 2).

Unexamined-Japanese-Patent No. 2010-257395 JP 2002-328134 A

  By the way, when determining the operation state of the object on which the information processing apparatus is mounted, the axes of the coordinate system of the information processing apparatus and the sensor, that is, the so-called local coordinate system and the global coordinate system do not coincide. It is necessary to carry out coordinate transformation for the

  Some information processing apparatuses such as mobile phone terminals include a main microcomputer that processes an operation on the information processing apparatus and executes communication, and a sub-microcomputer that processes output information from an acceleration sensor or the like. The main microcomputer is high-performance because it operates not only the control of the entire apparatus but also an application such as a browser. In general, it may be considered that the arithmetic processing performance of the microcomputer and the power consumption are in a proportional relationship. And in the information processing apparatus, the driving time (duration of the battery) is one of the important factors. If the power consumption in the information processing apparatus is reduced, the driving time of the information processing apparatus can be extended. Therefore, the main microcomputer is put to sleep as much as possible. Therefore, usually, a configuration is employed in which the main microcomputer does not constantly monitor the output of the acceleration sensor.

  Therefore, the sub-microcomputer of the information processing apparatus is used. Generally, the power consumption of the sub microcomputer is about 1/200 to 1/1000 of the main microcomputer. That is, even if the sub-microcomputer is always operated, there is no problem in terms of power consumption. However, the performance of the sub-microcomputer is less powerful than that of the main microcomputer, and it can be executed with simple operations such as integer arithmetic and memory operations. In order to constantly evaluate the motion of an object using a sub-microcomputer, necessary processing is performed based on the above simple calculation.

  For example, when converting values output in the coordinate system of the acceleration sensor into values in the so-called global coordinate system in the top, bottom, left, and right, such conversion can be performed by performing an operation using a rotation matrix. This conversion is usually performed using a trigonometric function. However, as described above, with a sub-microcomputer that can perform integer arithmetic operations and memory operation processing, it is not possible to perform operations such as trigonometric functions with high processing load.

  The technique of the present disclosure has been made in view of the above, and an object thereof is to measure the walking state of a tetrapod animal equipped with the measuring device while suppressing the power consumption in the measuring device. .

  In one aspect, the measurement program according to the present disclosure causes the measurement device attached to the tetrapod animal to acquire acceleration data of the three-dimensional space from the acceleration sensor attached to the tetrapod animal, and uses the acquired acceleration data The gait of the tetrapod is specified, and the gait of the identified tetrapod is used to determine whether or not the predetermined leg of the tetrapod has landed, and it is determined that the predetermined leg of the tetrapod has landed. The position information indicating the position of the quadruped animal when it was generated is generated using the position information and time information indicated by the GPS data acquired from the GPS sensor attached to the tetrapod animal, and the generated position information and GPS data are The average value of the acceleration in the movement of the four-legged animal from when it is determined that the predetermined leg of the four-legged animal has landed to when it is determined again that the predetermined leg of the four-legged animal has landed using the indicated time information. Calculate.

  According to the technology of the present disclosure, it is possible to measure the walking state of the tetrapod animal equipped with the measuring device while suppressing the power consumption of the measuring device.

FIG. 1 is a schematic configuration view showing a hardware configuration of a mobile phone terminal in one embodiment. FIG. 2 is a functional block diagram showing functional parts of the mobile telephone terminal in one embodiment. FIG. 3A is a diagram showing a vector of sensor output of an acceleration sensor in one embodiment. FIG. 3B is a diagram showing a vector of sensor output of an acceleration sensor in one embodiment. FIG. 4A is a schematic view showing the walking state of the user in one embodiment. FIG. 4B is a schematic view showing the walking state of the user in one embodiment. FIG. 5A is a diagram showing a vector of sensor output of an acceleration sensor in one embodiment. FIG. 5B is a diagram showing a vector of sensor output of an acceleration sensor in one embodiment. FIG. 6A is a diagram showing a vector of sensor output of an acceleration sensor in one embodiment. FIG. 6B is a schematic view showing a sagittal plane, a coronal plane, and a cross section of the user in one embodiment. FIG. 7 is a diagram showing a vector of sensor output of the acceleration sensor in one embodiment. FIG. 8 is a schematic configuration view showing a partial configuration of the sub-microcomputer in one embodiment. FIG. 9A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 9B is a schematic view showing the position and orientation of the user in one embodiment. FIG. 10 is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 11A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 11B is a graph showing temporal changes in angular velocity in one embodiment. FIG. 12A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 12B is a graph showing the relationship between the walking state of the user and the angular velocity in one embodiment. FIG. 13A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 13B is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 13C is a graph showing the relationship between the walking state of the user and the angular velocity in one embodiment. FIG. 14 is a diagram illustrating a horse's gait in one embodiment. FIG. 15 is a schematic block diagram showing a part of the configuration of the sub-microcomputer in one embodiment. FIG. 16A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 16B is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 16C is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 17 is a graph showing an example of an experimental result of gait determination of a horse according to an embodiment. FIG. 18 is a diagram illustrating the positional relationship between the horse body and the portable terminal in one embodiment. FIG. 19A is a schematic view showing a temporal change in the direction of the angular velocity in the vertical direction of the horse in one embodiment. FIG. 19B is a schematic view showing a temporal change in the direction of the angular velocity in the vertical direction of the horse in one embodiment. FIG. 20 is a schematic configuration view showing a configuration of a part of the sub-microcomputer in one embodiment. FIG. 21 is a graph showing the relationship between vertical movement and horizontal rotation of a horse according to an embodiment. FIG. 22A is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 22B is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 23 is a graph showing an example of an experimental result of the front determination of a horse in one embodiment. FIG. 24 is a schematic block diagram showing a part of the configuration of the sub-microcomputer in one embodiment. FIG. 25 is a graph showing the relationship between the state of each front leg of the horse and the vertical acceleration in one embodiment. FIG. 26 is a graph showing the relationship between each step of the horse and the vertical acceleration in one embodiment. FIG. 27 is a graph showing horse walking regularity and symmetry in one embodiment. FIG. 28 is a graph showing the correlation of horse walking regularity and symmetry in one embodiment. FIG. 29 is a diagram showing the relationship between each step of the horse and vertical acceleration and kinetic energy, and the relationship between kinetic energy and regularity and symmetry in one embodiment. FIG. 30A is a graph of vertical acceleration per step of a horse in one embodiment. FIG. 30B is a graph of vertical acceleration of each step of a horse in one embodiment. FIG. 31 is a schematic configuration view showing configurations of various devices in one embodiment. FIG. 32 is a functional configuration diagram showing functional parts of the measurement terminal in one embodiment. FIG. 33 is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 34 is a graph showing the correlation of horse walking regularity and symmetry in one embodiment. FIG. 35 is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 36 is a graph showing an example of the result of horse walking determination in one embodiment. FIG. 37A is a graph showing regular regions of regularity and symmetry set in one embodiment. FIG. 37B is a graph showing anomalous regions of regularity and symmetry set in one embodiment. FIG. 38 is a flowchart executed by the sub-microcomputer in one embodiment. FIG. 39 is a graph showing the relationship between regularity and symmetry measured in one embodiment. FIG. 40 is a schematic configuration view showing configurations of various devices in one embodiment. FIG. 41 is a functional configuration diagram illustrating functional units of the measurement terminal according to an embodiment. FIG. 42 is a graph of the vertical acceleration of each step of the horse in one embodiment. FIG. 43A is a graph showing the measurement results of the acceleration accompanying the movement of a horse in the prior art. FIG. 43B is a graph showing measurement results of acceleration accompanying movement of a horse in one embodiment. FIG. 44 is a graph of vertical acceleration of each step of a horse in one embodiment. FIG. 45 is a graph showing an abnormal area set in one embodiment. FIG. 46 is a flowchart executed by the sub-microcomputer in one embodiment.

First Embodiment
Hereinafter, the mobile phone terminal and the control method thereof in the first embodiment will be described with reference to the drawings. In this embodiment, it is assumed that a person (user) who is an exercise body carries a mobile phone terminal.

As shown in FIG. 1, the mobile phone terminal 10 in the present embodiment includes an antenna 101, an RF (Radio Frequency) unit 102, a main microcomputer 103, a sub microcomputer 104, and a button 10.
5, a display 106, a touch panel 107, a ROM (Read Only Memory) 108, a RAM (Random-Access Memory) 109, an acceleration sensor 110, and an angular velocity sensor 111. However, the technology disclosed in the present embodiment is not limited to the mobile phone terminal 10, and can be applied to information processing devices in general.

The mobile phone terminal 10 wirelessly communicates with an external wireless transmission / reception device (not shown) such as a base station via the antenna 101. The RF unit 102 modulates a baseband signal to a frequency band (RF band) at the time of transmission, and demodulates signals at these frequency bands to a baseband signal at the time of transmission / reception with an external wireless transmission / reception device.

  The main microcomputer 103 executes the control program 108 a stored in the ROM 108 to realize various functions of the mobile phone terminal 10. The sub-microcomputer 104 arithmetically processes sensor outputs of the acceleration sensor 110 and the angular velocity sensor 111, and converts it into an acceleration or an angular velocity of the coordinate system of the user of the mobile phone terminal 10, that is, the global coordinate system.

  The button 105 is a hardware key used by the user to operate the mobile phone terminal 10. The display 106 displays various information such as results of various processes executed in the mobile phone terminal 10. Further, the display 106 displays processing results, processing states, contents and the like of various processing executed by the main microcomputer 103 and the sub microcomputer 104. The touch panel 107 receives a touch operation or the like by the user. The touch panel 107 is disposed to overlap the display 106. The user performs a touch operation on the touch panel 107 with a finger or the like to select a display item such as an icon displayed on the display 106.

  The ROM 108 stores various programs executed in the mobile phone terminal 10. The RAM 109 is used as a temporary storage area when the main microcomputer 103 and the sub-microcomputer 104 execute various processes. That is, the RAM 109 is used as a work area, and stores specific information of the mobile phone terminal such as various programs and terminal numbers. In addition, the RAM 109 stores information of settings and programs input by the user, data provided to the user via the display 106, and the like.

  The main microcomputer 103 can execute high-speed arithmetic processing but consumes a large amount of power. Therefore, it is not preferable to keep the operating state at all times, so the operating time of the main microcomputer 103 is subject to a certain limit, and the operating state is temporarily shifted to the inactive state. By the way, when the user's operation state is determined, the user's operation data is constantly acquired, and the operation state is determined based on the acquired operation data.

  Therefore, in the present embodiment, the sub-microcomputer 104 is maintained in the operating state at least while the user's operation is occurring, receives the sensor output of the acceleration sensor 110 or the angular velocity sensor 111, and calculates the sensor output. Generate acceleration data of coordinate system. The sub-microcomputer 104 executes processing with low processing load such as simple arithmetic processing such as arithmetic operations and access processing to memories such as the ROM 108 and the RAM 109.

  Therefore, the main microcomputer 103 is activated when necessary, and the sub-microcomputer 104 converts the statistical information into a value easy to be understood by the user based on the determination result of the user's operation state constantly or for a long time Then, the generated information is displayed on the display 106. The generated information may be transmitted to the outside via the antenna 101.

  In the present embodiment, the application or the like stored in the ROM 108 of the mobile phone terminal 10 is executed by the CPU 103. By executing such an application, as shown in FIG. 2, the mobile phone terminal 10 functions as an inner product value calculation unit 201, an outer product calculation unit 202, an absolute value calculation unit 203, and an operation state specification unit 204. However, any one or more of the inner product value calculating unit 201, the outer product calculating unit 202, the absolute value calculating unit 203, and the operation state specifying unit 204 may be a hardware circuit.

Next, a configuration, an algorithm, and a process of determining a predetermined operation performed by the user, which is executed in the sub-microcomputer 104 of the mobile phone terminal 10 in this embodiment, will be described.

FIGS. 3A and 3B show an example of a method of extracting acceleration information in the vertical direction (vertical direction) in the coordinate system of the user from the sensor output of the acceleration sensor when the posture of the user carrying the mobile phone terminal is unknown. FIG. Here, it is assumed that the user is walking. As shown in FIG. 3A, the X axis, the Y axis, and the Z axis orthogonal to one another constitute a local coordinate system of the mobile phone terminal. In FIG. 3A, the sensor output A of the acceleration sensor is given by A = (x, y, z) ^T. In addition, T means a transposed matrix. And as shown to FIG. 3A, the sensor output A of an acceleration sensor can be considered to be the synthesis | combination of the vibration S and the gravitational acceleration G accompanying a user's walk.

  Here we outline human walking. FIG. 4A and FIG. 4B are diagrams schematically showing movements of the torso and feet in human walking. FIG. 4A is a view of human walking as viewed from directly above (vertically upward), and FIG. 4B is a view of human walking as viewed from the side (horizontal direction). The trunk 20 corresponds to the human hip to head portion, and the legs 20L and 20R correspond to the left leg and the right leg, respectively. As shown in FIGS. 4A and 4B, during walking, the left and right legs 20L and 20R alternately come to the ground while walking, and the trunk 20 sways in the vertical and horizontal directions and the front and back directions. Also, as shown in FIGS. 4A and 4B, the period of vertical movement of the trunk 20 is a period in which one of the left and right legs 20L, 20R leaves the ground again and arrives again, and the other leaves the ground again. Match with On the other hand, the left and right swings of the trunk 20 sway in the left direction of the human while the left leg 20L is on the ground, and sway in the right direction of the human while the right leg 20R is on the ground. Therefore, the period of swinging of the left and right of the trunk 20 at the time of human walking is twice the period of the period of vertical movement, that is, the frequency of 1⁄2.

  In addition, during walking, when the user lands the left leg 20L or the right leg 20R on the ground, the downward acceleration is greatest. Then, after landing on the left leg 20L or the right leg 20R, an upward acceleration occurs when the trunk 20 is raised, and the upward acceleration is minimized at the highest point reached by the trunk 20, and thereafter, it is replaced Downward acceleration occurs. Further, with regard to acceleration in the left and right direction, when the left leg 20L is landed, acceleration in the left direction is generated, and when the right leg 20R is landed, acceleration in the right direction is generated.

Returning to FIGS. 3A and 3B, as described with reference to FIGS. 4A and 4B, it can be said that the vibration S accompanying the user's walking is a repetitive motion of a predetermined cycle. Therefore, when the sensor output A of the acceleration sensor is passed through a low pass filter (LPF) or high pass filter (HPF) of an appropriate cutoff frequency, the sensor output A can be separated into the vibration component vector S and the gravity acceleration vector G. it can. The acceleration in the vertical direction accompanying walking can be represented by a vector V parallel to the gravitational acceleration vector G in the vibration component vector S. Assuming that the magnitude of the vector V is v and the angle between the vibration component vector S and the gravity acceleration vector G is θ, v is given by the following equation.
v = | S | cos θ (1)

Here, the inner product of the vibration component vector S and the gravity acceleration vector G is considered. The inner product is obtained by multiplying the x, y and z components of the vector and then summing the multiplied components. The inner product of the vibration component vector S and the gravity acceleration vector G is given by the following equation from the definition.
S · G = | S | | G | cosθ (2)

By applying equation (2) to equation (1), the following equation (3) is obtained.
v = S · G / | G | (3)

  Here, the magnitude | G | of the gravitational acceleration vector G is substantially constant on the earth. That is, by passing the LPF and HPF to the sensor output A of the acceleration sensor and taking the inner product of the outputs of the respective filters, the value proportional to the acceleration in the upward direction accompanying the user's walking motion and the acceleration in the downward direction were proportional You can get numbers and numbers. In the above description, it is assumed that the user is walking. However, even when the user performs an operation other than walking, a vibration component vector and a gravity acceleration vector can be obtained by passing the LPF and the HPF to the sensor output A of the acceleration sensor based on the above algorithm.

  Next, FIGS. 5A and 5B show a method of extracting acceleration information in the horizontal direction (left and right direction) in the coordinate system of the user from the sensor output of the acceleration sensor when the posture of the user carrying the mobile phone terminal is unknown. It is a figure which shows an example. Here also, it is assumed that the user is walking. Also, it is the same as FIGS. 3A and 3B that the sensor output A of the acceleration sensor is considered to be a combination of the vibration component vector S and the gravity acceleration vector G accompanying the user's walking. Here, assuming that the vibration component vector S is the sum of the vector V parallel to the gravity acceleration vector G and the vector H orthogonal to the vector V, the vector H corresponds to the horizontal direction of the vibration component vector S accompanying the user's walking. It can be considered as a component of

Here, in order to obtain the magnitude | H | of the vector H, the outer product of the vibration component vector S and the gravity acceleration vector G is considered. Similar to the inner product, the outer product is also obtained by arithmetic operations, so that it can be calculated by a microcomputer having a processing performance lower than that of the main microcomputer 103, such as the sub microcomputer 104. As shown in FIG. 5B, the outer product of the vibration component vector S and the gravitational acceleration vector G has a size equal to the area Z of the parallelogram formed by the vectors S and G in the direction orthogonal to the vectors S and G by definition. Is a vector of That is, the outer product of the vibration component vector S and the gravity acceleration vector G is given by the following equation (4).
Z = | S × G | (4)

As shown in FIG. 5B, the area Z of the parallelogram can also be described as | G | × | H |. If this is applied to equation (4), the following equation (5) is obtained.
| H | = Z / | G | = | S × G | / | G | (5)

  As mentioned above, | G | is almost constant on the earth. Therefore, the magnitude of the horizontal component of the vibration associated with walking can be obtained by calculating the absolute value of the sensor output A of the acceleration sensor by taking the outer product of the vibration component vector obtained by passing the LPF and HPF and the gravity acceleration vector. You can get a value proportional to

  Next, FIG. 6A is an example of a method of extracting acceleration information of each direction in the user's coordinate system from top, bottom, left, right, and back from the sensor output of the acceleration sensor when the posture of the user carrying the mobile phone terminal is unknown. FIG. Here also, it is assumed that the user is walking. In this method, it is assumed that the vibration component vector S is vertically and horizontally separated by the method shown in FIGS. 5A and 5B. In the present method, it is used that when the user's walking enters a steady state (constant velocity movement), almost no acceleration in the longitudinal direction of the user's trunk occurs. When it is considered that acceleration does not occur in the front-rear direction of the user's trunk, horizontal acceleration | H | obtained in FIGS. 5A and 5B indicates the magnitude of acceleration in the left-right direction.

Furthermore, as shown in FIG. 6B, considering so-called sagittal plane, coronal plane, and cross section of the user's torso, the vibration component vector S exists on the coronal plane of the user's torso. Therefore, the outer product F of the vibration component vector S and the gravity acceleration vector G faces the front direction (front side) or the back direction (back side) of the user's trunk. The front-rear direction of the outer product F depends on the left-right direction of the vibration component vector S. Here, assuming that the vector K pointing in the forward direction of the user's torso is known, it can be seen that the vector F points in the forward direction if the inner product of the vectors F and K is positive. Also, when the inner product of the vectors F and K is negative, it can be seen that the vector F is directed backward. In addition, when the vector F turns to the front, the vibration component vector S turns to the right (right hand side) of the user's trunk. In addition, when the vector F faces in the backward direction, the vibration component vector S faces in the left direction (left hand side) of the user's trunk.

Furthermore, it is possible to distinguish between the right direction and the left direction for the acceleration in the left and right direction of the user's torso according to the following equation (6).
h = sign (F · K) | H | (6)
Here, “sign (F · K)” indicates the sign of the inner product of the vector F and the vector K. Therefore, by adding the sign of the inner product to | H | from Equation (6), it can be determined whether the vector H is the acceleration vector in the right direction or the acceleration vector in the left direction of the user's torso.

  Next, in FIG. 7, when the posture of the user carrying the mobile phone terminal is unknown, from the sensor output of the acceleration sensor, the acceleration information of each direction in the top, bottom, left, right, front and back in the global coordinate system of the user is specified specifically. It is a figure which shows an example of the method of taking out with a value. Here also, it is assumed that the user is walking. In the example shown in FIGS. 6A and 6B, it is assumed that acceleration in the front-rear direction does not occur in the steady state. However, in the example shown in FIG. 7, by using the high resolution acceleration sensor, it is also possible to detect minute fluctuations in the longitudinal acceleration.

Here, as shown in FIG. 4A and FIG. 4B, the longitudinal fluctuation is synchronized with the vertical fluctuation. And when accelerating downward, it is accelerating forward, when accelerating upward, it is accelerating backward (it may be considered as lower front and upper back). Furthermore, the cycle of fluctuation in the lateral direction is twice as large as the fluctuation cycle in the vertical direction and the longitudinal direction. Therefore, utilizing this characteristic, in the example shown in FIG. 7, the sensor output of the acceleration sensor using the band pass filter is represented by the vector S ₁ mainly composed of the longitudinal acceleration and the vertical acceleration and the lateral acceleration. some separating the vector S ₂ containing the longitudinal acceleration. Furthermore, the acceleration in the front-rear direction and the acceleration in the left-right direction are extracted by an appropriate outer product calculation in a state in which both positive and negative are unknown. The acceleration in the front-rear direction and the acceleration in the left-right direction can be determined by using the phase difference between the extracted acceleration vector in the front-rear direction and the acceleration vector in the left-right direction with the acceleration vector in the vertical direction.

  The method of using the sensor output of the above acceleration sensor separates the acceleration information obtained from the acceleration sensor into acceleration information in each axial direction in the global coordinate system in a state where the posture of the user carrying the mobile phone terminal is unknown. By the way, as the operation information obtained by the sensor, there is an angular velocity other than the acceleration. The angular velocity information can be obtained by using an angular velocity sensor (gyro). However, as for the angular velocity information obtained from the sensor output of the angular velocity sensor, it is unclear from the sensor output itself about which axis in the global coordinate system the value is, as in the above-mentioned acceleration information. Therefore, the sensor output of the angular velocity sensor is converted into the angular velocity around the so-called pitch axis, roll axis, and yaw axis of the user. Generally, in the terminal, the detection axis of the acceleration sensor and the detection axis of the angular velocity sensor are attached so as to coincide with each other, so it is assumed in the present embodiment that the detection axes coincide with each other. If the respective detection axes do not coincide with each other, the deviation of the two detection axes is measured beforehand by an appropriate calibration means, and a coordinate transformation matrix corresponding to the measured deviation is applied to the output value of one of the sensors. Each sensor output can be treated as if the detection axes coincide.

The sensor coordinate system 11 of the angular velocity sensor in the mobile phone terminal is set to three mutually orthogonal XYZ axes, and sensor output values of angular velocity sensors in the directions of X, Y and Z axes in the sensor coordinate system 11 Let xr, yr and zr respectively. In addition, with respect to the coordinate system of the user carrying the mobile phone terminal, so-called global coordinate system 22, three axes of pitch axis, roll axis, and yaw axis orthogonal to one another are set, and pitch axis roll Y axis in global coordinate system 22, yaw The sensor output values of the angular velocity sensor in each axial direction of the axes are p, r, and y, respectively.

Then, the sensor output value SL (= (xr, yr, zr) ^T ) of the angular velocity sensor is converted into the angular velocity SG (= (p, r, y) ^T ) around the pitch, roll and yaw axes of the user. What is done is equivalent to performing coordinate transformation from the sensor coordinate system 11 to the global coordinate system 22. Here, for conversion from SL to SG, Y = RX may be calculated using coordinate conversion matrices R of the coordinate systems 11 and Σ2. Here, R is a 3 × 3 matrix and is represented by the following matrix.

Each element a00 to a22 of the coordinate conversion matrix R can obtain relative angles about each axis of the two coordinate system through an operation using an appropriate trigonometric function. In addition, since this operation is known, the detailed description is omitted here. However, as described above, it is not desirable to perform operations such as trigonometric functions with the sub-microcomputer. Therefore, the sensor coordinate system Σ2 of viewed .SIGMA.1 XYZ orientation units each vector _e x of each _axis, e y, and expressed in _{e z.} Then each element of the vector _{e x, e y, e z} , e x = using the elements of the coordinate transformation matrix _{R (a00, a01, a02)} T, e y = (a10, a11, a12) T, e _z = (a20, a21, a22) It can be expressed as ^T. That is, if the vector e _x, e _y, even found e _z, the value of the sensor output of the angular velocity sensor, it is possible to convert the pitch axis in the global coordinate system, a roll shaft, each angular speed of the yaw axis.

  FIG. 8 shows a part of the circuit of the sub-microcomputer 104 of the mobile phone terminal 10 in the present embodiment. As shown in FIG. 9A, the sensor output A of the acceleration sensor 110 is input to the LPF 104c and the HPF 104d. As described above, the gravitational acceleration vector G is output from the LPF 104c, and the vibration component vector S is output from the HPF 104d.

  Further, as shown in FIG. 8, the sub-microcomputer 104 includes an outer product calculation processing unit 104 e, a multiplication processing unit 104 f, an LPF 104 g, a sign determination unit 104 h, and an inner product calculation processing unit 104 i. The outer product calculation processing unit 104e obtains the outer product of the vector G output from the LPF 104c and the vector S output from the HPF 104d. The multiplication processing unit 104 f multiplies the outer product S × G by the coefficient of +1 or −1. Further, the LPF 104 g obtains an average value of the output of the multiplication processing unit 104 f. Since the outer product S × G is a vector in the front-rear direction, when multiplying the coefficient +1 or −1 by the multiplication processing unit 104 f, it is possible to obtain a vector in one direction in the front-rear direction. Therefore, the LPF 104g outputs a vector in one of the front and rear directions. Then, the inner product calculation processing unit 104i obtains the inner product of the vector in one direction in the front-rear direction and the outer product S × G output from the outer product calculation processing unit 104e.

In the inner product calculation processing unit 104i, the sign of the inner product is +1 if the outer product S × G is the forward direction (ie, the vibration component vector S is the right direction), and the outer product S × G is the backward direction (ie, the vibration component vector S) Is left direction), it becomes -1. Therefore, by detecting this code by the code determination unit 104h and sending +1 or -1 to the multiplication processing unit 104f, a vector H indicating the forward direction is obtained from the LPF 104g. Here, for convenience, the vector H is assumed to be a vector indicating the “front” direction, but depending on the setting of initial values in the loop of the multiplication processing unit 104 f, the LPF 104 g, the sign determination unit 104 h, and the inner product calculation processing unit 104 i, The "front" direction described in may also indicate a "back" direction. In the following description, the direction of the vector H output from the LPF 104 g is assumed to be the forward direction on the roll axis (Y axis) of the global coordinate system 22.

The direction of the vector G output from the LPF 104 c is equal to the yaw axis (Z axis) direction of the global coordinate system 22. The inner product calculation processing unit 104 j calculates an inner product of the vector G output from the LPF 104 c and the value X of the sensor output of the angular velocity sensor 111. The inner product is given by the following equation (7).
G · X = | G | e _z · X = gy (7)

  Here, g is the magnitude of gravitational acceleration, which is substantially constant on the ground. Therefore, the output value of the inner product calculation processing unit 104 j is equal to a value obtained by multiplying the angular velocity of rotation about the yaw axis by a constant. Therefore, the angular velocity around the yaw axis is calculated by equation (7) in the inner product calculation processing unit 104 j.

Next, the forward vector H can be expressed as Expression (8) using a forward unit vector e _y viewed from the sensor coordinate system and an appropriate real number k ₁ .
H = k ₁ e _y (8)
Since | H | depends on the direction and size of the vector S, k ₁ is not a constant value. However, by making the time constant of the LPF 104g sufficiently large, the magnitudes of the vectors H outputted from the LPF 104g can be regarded as being averaged and becoming a constant value in a predetermined short time. When the inner product of the vector H and the value X of the sensor output of the angular velocity sensor 111 is calculated, the following equation (9) is obtained. In the present embodiment, the inner product calculation processing unit 104m performs the calculation of equation (9).
H · X = k ₁ e _y · X = k ₁ r (9)
That is, in a short time, it is considered that the inner product value obtained by the equation (9) is proportional to the magnitude of the angular velocity around the roll axis in the global coordinate system 22. However, although it is possible regarded as k ₁ is approximately constant value, the magnitude of the actual angular velocity unknown. Therefore, in the present embodiment, the user's operation is determined by using the positive and negative of the angular velocity around the roll axis as a value related to the angular velocity. Details will be described later.

Further, considering the outer product K of the forward direction vector H and the gravitational acceleration vector G, the outer product K is a vector indicating the right direction or the left direction in the global coordinate system. In the present embodiment, the outer product K is calculated by the outer product calculation processing unit 104k. Here, when the inner product of the outer product K and the value X of the sensor output of the angular velocity sensor 111 is obtained, the following equation (10) is obtained. In the present embodiment, the inner product calculation processing unit 1041 performs the calculation of equation (10).
_{K · X = k 2 e x} · X = k 2 p ··· (10)
Therefore, the inner product value obtained by the equation (10) can take out a value related to the angular velocity around the pitch axis in the global coordinate system 22.

  Next, in the present embodiment, the cellular phone terminal 10 is calculated using the angular velocity around the yaw axis, the information related to the angular velocity around the roll axis, and the information related to the angular velocity around the pitch axis calculated in the sub-microcomputer 104. An example of processing for determining various operations of the user carrying will be described.

FIG. 9A shows a flowchart of processing executed by the sub-microcomputer 104 in an example in which indoor navigation is performed using the angular velocity ωy around the yaw axis obtained by the circuit shown in FIG. In the following description, step is abbreviated as S. What is important in indoor navigation is to know the relative position (x, y) of the user U at XY coordinates with an appropriate origin set as the origin O, as shown in FIG. 9B. If the position of the user U is known, it can be known where in the house, and an appropriate navigation service can be provided. 9A.
The processing of S101 to S109 of the flowchart shown in FIG. 5 is a loop that is repeatedly called for each appropriate sampling time (ΔT). In addition, the said sampling time is suitably set so that 1 loop may be completed in a short cycle with respect to a user's walk period.

  In S101, the sub-microcomputer 104 obtains each speed ωy around the yaw axis of the user. Next, the process proceeds to S102. In S102, the sub-microcomputer 104 confirms whether or not the direction in which the user U is facing forward can be identified by a known means. This is performed by, for example, detecting that the user U has entered or left the room from a specific door indoors or that the user U is walking in the corridor using an external device such as an RFID, and matching the detection result with the map information Thus, it is possible to estimate which direction the user U is currently facing, that is, the current orientation θ shown in FIG. 9B. If the estimation is successful (S102: Y), the sub-microcomputer 104 regards the estimated orientation as the current orientation θ (S103). If the estimation fails (S102: N), the sub-microcomputer 104 adds the integral value of the angular velocity in the above sampling period ΔT to the current direction θ (S104). When the processes of S103 and S104 end, the process proceeds to S105.

  In S105, the sub-microcomputer 104 calculates the current number of steps of the user U. In the process of S105, the cellular phone terminal 10 acquires the calculation result of the pedometer which is an external device, and the sub-microcomputer 104 utilizes the calculation result. Then, the process proceeds to S106. In S106, as in S102, it is confirmed whether the current position of the user U can be estimated by known means. If the user's current position has been successfully estimated (S106: Y), the process proceeds to S107 to consider the estimated position as the current position. If the estimation of the current position of the user has failed (S106: N), the process proceeds to S108, and the sub-microcomputer 104 determines whether or not the number of steps of the pedometer for which the calculation result has been obtained in S105 has been counted up. When the sub-microcomputer 104 determines that the pedometer has been counted up (S108: Y), the sub-microcomputer 104 updates the current position (x, y) based on the movement distance L taken by one step of the user U (S109). Here, the movement distance L may be calculated based on the height or the like of the user U, or may be appropriately estimated using the difference between the estimated position and the current position in the process of S107. Therefore, even if the user's azimuth estimation (S102) and position estimation (S106) are not successful, the azimuth and position can be estimated using the angular velocity around the user's yaw axis by the processing of S104 and S109.

  Next, FIG. 10 shows a flowchart for determining whether the user is pushing the cart, currently walking or not walking. Generally, when a person pushes a cart such as a carriage or a stroller, both hands are fixed to the carriage and moved, so that the person's rotation around the yaw axis, that is, the angular velocity tends to be smaller compared to normal walking. There is. Therefore, it is determined by the processing shown in FIG. 10 whether the user is pushing the cart, currently walking or not walking.

  In S201, the sub-microcomputer 104 acquires the respective speeds ωy around the yaw axis of the user. Next, the process proceeds to S202. In S202, the sub-microcomputer 104 determines whether the user is currently walking using the acceleration sensor 110 or the like. This determination process is obtained by using a calculation result of an external pedometer or the like, and is a well-known process, so detailed description will be omitted. When the sub-microcomputer 104 determines that the user is not currently walking (S202: N), the sub-microcomputer 104 sets the value cart of the pushing probability of cart to 0 (S210), and determines that the user is not walking (S211).

On the other hand, when the sub-microcomputer 104 determines that the user is walking in S202 (S202: Y), the sub-microcomputer 104 determines whether the absolute value of the angular velocity ωy around the yaw axis is larger than a predetermined threshold. If the absolute value of the angular velocity ωy is larger than the predetermined threshold (S203: Y), the sub-microcomputer 104 increases the probability of pushing the truck by N1 (S204). In S205 and S206, the sub-microcomputer 104 performs processing to limit the truck pushing accuracy not to be larger than a predetermined value N2. Furthermore, in S207 and S208, the sub-microcomputer 104 decrements the truck pressing probability when the truck pressing probability has a value other than zero. Then, the sub-microcomputer determines whether the truck pressing accuracy is greater than a predetermined threshold N3 (S209), and determines that the truck pressing is performed (S209: Y) if the truck pressing accuracy is greater than N3 (S213), If the probability of pushing the dolly is N3 or less (S209: N), it is determined that the user is walking (S212). Whether the user is pushing the dolly or is currently walking by changing the value of the probability of pushing the doll within a predetermined range (0 to N2) based on the angular velocity ωy around the yaw axis of the user by the above processing It can be determined whether you are walking.

  Next, FIG. 11A shows a flowchart for diagnosing the walking state of the user, specifically, the balance between the left and right of the body at the time of walking. In the execution of this flowchart, it is determined in advance whether the user is walking or not, and the sub-microcomputer 104 executes the flowchart if it is determined that the user is walking. Further, this flowchart is repeatedly executed at a predetermined sampling interval.

  In S301, the sub-microcomputer 104 calculates information related to the angular velocity ωr around the roll axis of the user as shown in FIG. Here, “related” specifically means, as described above, that the value of the angular velocity ωr is not the actual angular velocity around the roll axis but a constant multiple of the actual angular velocity around the roll axis is there. In the following, for convenience, the information related to the angular velocity ωr will be described as the angular velocity ωr. Next, the process proceeds to S302.

  In S302, the sub-microcomputer 104 determines whether the product of the current angular velocity ωr and the angular velocity ωr 'one sampling before is negative. Although ω r ′ is a value substituted in S 310 described later, an initial value of ω r ′ can be set as appropriate. The fact that ωr × ωr ′ is negative in the process of S302 means that the graph showing the time change of ωr crosses zero (the sign of the value of ωr changes) as shown in FIG. 11B. Therefore, as shown in FIG. 11B, the sum of the half cycle of ωr, that is, the numerical value associated with the movement amount from the center is sum, and while the graph does not cross zero (S302: N), continues to add ωr to sum (S302) S309).

  As shown in FIG. 11B, depending on the cycle, sum may be integrated as a negative value, so when the graph crosses zero (S302: Y), the sub-microcomputer 104 makes the value of sum positive in S303. After that, the balance is calculated by taking the ratio to the total sum 'of half cycle before (S305, S306). The sub-microcomputer 104 determines the magnitudes of sum and sum 'in S304, and divides the large value by the small value to calculate the balance always as a value larger than one. Therefore, it can be said that the balance between the left and right of the body at the time of walking is better as the calculated value approaches 1.

  In this flowchart, after the balance is calculated, the current total sum sum is substituted into the pre-sum total sum 'for the next balance calculation (S307), and ωr obtained in S301 is substituted into the current total sum (S308) . By the above processing, by utilizing the period of time change of information related to the angular velocity around the roll axis of the user, it is possible to diagnose the balance between the left and right of the body at the time of walking of the user. In the above-described processing example, the sum for a half cycle is obtained, but the time range for obtaining the sum is not limited to a half cycle, and can be appropriately determined. Also, in the above processing example, the balance between the left and right of the body at the time of walking is calculated based on the ratio of the sum, but as shown in FIG. 11B, the ratio (h / h ') of the maximum value of the graph and the time The balance may be calculated based on the ratio (t / t ′) of

Next, FIG. 12A shows a flowchart for determining whether or not an abnormal state such as becoming nearly tipped when the user walks has occurred. In the execution of this flowchart, it is determined in advance whether the user is walking or not, and the sub-microcomputer 104 executes the flowchart if it is determined that the user is walking. Further, this flowchart is repeatedly executed at a predetermined sampling interval.

  FIG. 12B aligns the x-axis in the local coordinate system of the mobile phone terminal 10 with the pitch axis in the global coordinate system of the user, and changes with time the angular velocity around the x-axis when the user walks. Is a graph showing As shown in the figure, when an abnormal state occurs, that is, when the user is about to fall, it can be seen that the angular velocity shows a characteristic change as compared to the change of the angular velocity in the walking state until then. Therefore, by executing the flowchart shown in FIG. 12A, the sub-microcomputer 104 can determine whether an abnormal state has occurred while the user is walking. 12B aligns the x-axis in the local coordinate system of the mobile phone terminal 10 with the pitch axis in the global coordinate system of the user for convenience of explanation, but as described with reference to FIG. The process of this flowchart can be normally executed without aligning the axes.

  In S401, the sub-microcomputer 104 calculates information related to the angular velocity ωp around the pitch axis of the user, as shown in FIG. Here, “related” specifically means, as described above, that the value of the angular velocity ωp is not the actual angular velocity around the pitch axis but a constant multiple of the actual angular velocity around the pitch axis is there. In the following, for convenience, information related to the angular velocity ωp will be described as the angular velocity ωp. The process then proceeds to S402.

  In S402, the sub-microcomputer 104 takes the absolute value ωp 'of ωp calculated in S401. Next, the process proceeds to S403. In S403, the sub-microcomputer 104 calculates the result ψ obtained by passing the low-pass filter (not shown) to ωp ′ calculated in S402. By setting the time constant of the low-pass filter to be sufficiently longer than the walking cycle, an amplitude value of angular velocity around the pitch axis averaged over a time range determined by the time constant can be obtained. The process then proceeds to S404. In S404, the sub-microcomputer 104 determines whether ωp ′ calculated in S402 is larger than a value obtained by multiplying the amplitude value ψ of the average angular velocity calculated in S403 by k (k is a positive integer). If ωp ′ is larger than k times ((S404: Y), the sub-microcomputer 104 determines that an abnormal state has occurred (S405). If ωp ′ is equal to or less than k times ψ (S404: N), the sub-microcomputer 104 ends the processing of this flowchart. By the above processing, it is possible to track the time change of the angular velocity around the pitch axis of the user, detect when the acceleration shows an abnormal change, and regard the user's walking in an abnormal state.

  Next, FIG. 13A and FIG. 13B show a flowchart for determining whether the user is walking while lifting a large load such as a trunk case with one hand. Note that “1” and “2” in the flowchart of FIG. 13A are continuous with “1” and “2” in the flowchart of FIG. 13B, respectively. FIG. 13C is a graph showing a temporal change in angular velocity around a roll axis in the user's global coordinate system when the user is carrying a package. As shown in the figure, the balance of the angular velocity around the roll axis becomes worse when carrying a load. Specifically, it appears in the time required for a half cycle and the maximum amplitude. In the figure, the time taken for one half cycle is shorter than the time taken for the other half cycle, and its maximum amplitude is large ("short and large"), and the other half cycle has its maximum amplitude at one half cycle. The maximum amplitude is decreasing ("long and small").

Therefore, in this flowchart, by evaluating the balance of the angular velocity around the roll axis, it is determined whether the user lifts the package with one hand and walks. In the execution of this flowchart, it is determined in advance whether the user is walking or not, and the sub-microcomputer 104 executes the flowchart if it is determined that the user is walking. Further, this flowchart is repeatedly executed at a predetermined sampling interval.

  In S501, the sub-microcomputer 104 calculates information related to the angular velocity ωr around the roll axis of the user as shown in FIG. Here, “related” specifically means, as described above, that the value of the angular velocity ωr is not the actual angular velocity around the roll axis but a constant multiple of the actual angular velocity around the roll axis is there. In the following, for convenience, the information related to the angular velocity ωr will be described as the angular velocity ωr. Next, the process proceeds to S502.

  In S502, as in the case shown in FIGS. 11A and 11B, it is determined whether or not the graph indicating the time change of ωr crosses zero (the sign of the value of ωr changes). When the zero crossing has not occurred (S502: N), the value cnt of the counter is incremented (S503). The value of cnt is set to 0 in S513 when a zero crossing occurs. After the process of S503, the sub-microcomputer 104 compares the current value of the angular velocity ωr with the maximum value or the minimum value of the angular velocity ωr after the start of the present flowchart in S504 to S507 to obtain the maximum value of the angular velocity ωr. Find the value and the minimum value. In S508, the sub-microcomputer 104 sets the value of the current angular velocity ωr to the value of the angular velocity ωr 'one sampling before.

  In S502, when the graph indicating the time change of ωr crosses zero (S502: Y), the process proceeds to S509. In S509 to S511, the ratio of the counter value cnt 'of the counter one half period before the time change of ωr to the value cnt of the current counter is calculated as the period balance value carry1. In FIG. 13B, a large value among the values of the counter is divided by a small value, and the value of carry1 becomes a value larger than 1, so the closer the value of carry1 is to 1, the balance between the left and right of the body in the user's walking I understand that it is good. That is, by calculating the balance of the cycles, it can be understood, in the graph shown in FIG. 13C, whether or not the time taken for the half cycle is alternately repeated.

  In S512 and S513, after replacing the value cnt 'of the counter half cycle before with the value of the current counter, the value cnt of the present counter is returned to the initial value 0. The process then proceeds to S514. As mentioned above, ωr takes positive and negative values alternately every half cycle. Further, the maximum value and the minimum value obtained in S504 to S507 indicate one of the peaks of the amplitude of ωr in the half cycle immediately before the occurrence of the zero crossing, and the other shows the amplitude of ωr in the half cycle one cycle before the half cycle. Indicates the peak of Therefore, in S514 to S516, the ratio of the maximum value to the minimum value is calculated as the value balance carry2 of the amplitude. In S517 to S519, the maximum value or the minimum value indicating the peak of the amplitude of ωr in the half cycle immediately before the half cycle is specified, and the value is set to 0 for the next calculation of carry2.

  Each of carry1 and carry2 obtained by the above process represents the balance of time change of ωr. Then, in the case of carry1, the threshold Th1 is compared with the threshold Th2 in the case of carry2, and when one or both of carry1 and carry2 takes a value larger than the threshold, the balance is a bad balance, that is, the user is carrying luggage. It is determined that there is a case (in this example, the package is lifted with one hand and walked) (S520, S521). By the above processing, it is determined whether or not luggage is being transported by evaluating the balance of the walking of the user based on the time width of each period of the angular velocity around the roll axis of the user and the magnitude of the amplitude in each period. Can.

When it is determined whether or not the package is being transported using either one of the carry1 and the carry2, it is not necessary to carry out the arithmetic processing relating to the other. For example, in the case of evaluating carry1 in S520, the process of obtaining the value of carry2 (S504 to S507, S514 to S519) can be omitted. Conversely, when evaluating carry2 in S520, ca
The process of obtaining rry1 (S503, S509 to S513) can be omitted. Furthermore, in the above processing example, when it is determined in S520 that the values of carry1 and carry2 are larger than the threshold, it is determined in S521 that the user is carrying a package, but as shown in S203 to S213 of FIG. The concept of accuracy can be used to determine whether or not the package is being transported.

Second Embodiment
Next, a second embodiment will be described with reference to the drawings. Also in the present embodiment, the portable terminal 10 'having the same configuration as the mobile phone terminal 10 of the first embodiment is used, and in the following description, the same configuration as the first embodiment has the same reference numerals. Detailed explanation is omitted by using. In a second embodiment, the gait of a tetrapod, for example a horse, is determined. The gait, that is, the way of running and walking changes depending on the moving speed of the four-legged animal such as a horse. FIG. 14 shows representative gaits of horses, (1) walk, (2) trot, (3) canter, and (4) gallop. In FIG. 14, the left side of the figure schematically shows the positional relationship of the leg of the horse of each gait, and each leg (right front (leg), left front (leg) in a period indicated by “1” to “8” of each gait. The presence or absence of grounding on the right rear (leg) and the left rear (leg) is shown on the right of each schematic diagram. The horse's gait may also include waves, paces, bounces, etc. However, since it is not so common as the 4 gaits shown in the figure, it will not be described here.

  Here, as can be seen from FIG. 14, the horse body floats completely in the periods shown by "4" and "8" of Gallop, during the periods shown by "7" to "8" of the canter. Also, in FIG. 14, in the trot, it appears that all four legs are in contact with the ground during the periods shown as “1” and “4”, but if actual data is confirmed in detail, leg switching is quick in these periods It is practiced and it is known that there is a time when the horse is floating in the air.

  That is, the fact that the horse is floating in the air means that the acceleration sensor attached to the horse is in a state of weightlessness. Therefore, when the horse is moving, when the acceleration sensor detects zero gravity, the horse can be considered to be in the trot, canter or gallop state. Therefore, in the present embodiment, the mobile phone terminal 10 is attached to the horse's body, and the sub-microcomputer 104 determines the horse's gait using the sensor output of the acceleration sensor 110.

  FIG. 15 shows the configuration of part of the sub-microcomputer 104 in the present embodiment. The sub-microcomputer 104 includes an LPF 104n, an HPF 104o, and an inner product calculation processing unit 104p, 104q. The sensor output of the acceleration sensor 110 is input to the LPF 104 n, the HPF 104 o, and the inner product calculation processing unit 104 q. The outputs of the LPF 104 n and the HPF 104 o are input to the inner product calculation processing unit 104 p. The series of processes performed by the LPF 104 n, the HPF 104 o, and the inner product calculation processing unit 104 p are the same as those in the first embodiment, and thus the description thereof is omitted. The gravitational acceleration component vector G ′ ′ and the vibration component vector S ′ ′ are respectively output from the LPF 104 n and the HPF 104 o. Therefore, the magnitude of the vibration component vector in the vertical direction in the horse's global coordinate system is output from the inner product calculation processing unit 104p. On the other hand, the inner product calculation processing unit 104 q performs inner product calculation on the sensor output itself of the acceleration sensor 110, that is, calculates the square of the absolute value of the sensor output. For convenience, it is assumed that the value of the output obtained from the inner product operation processing unit 104p is α, and the value of the output obtained from the inner product operation processing unit 104q is β.

FIGS. 16A to 16C show flowcharts for determining the gait of a horse using the values α and β obtained by the above process. Note that this flowchart is repeatedly executed at predetermined intervals using the values α and β obtained by the processing shown in FIG. In S601, the sub-microcomputer 104 determines whether or not the user is walking based on the output value α from the inner product arithmetic processing unit 104p. The determination process is well known and thus the description thereof is omitted. If it is determined that the horse is walking (
S601: Y), the process proceeds to S602. If it is determined that the horse is not walking (S601: N), the process proceeds to S619. In S619, the sub-microcomputer 104 determines that the horse is not walking after setting the accuracy n_trot of the trot, the accuracy n_run of the traveling, and the accuracy n_gallop of the gallop to 0 (S620). Here, the running accuracy n_run is a value used to determine that the current horse's gait is either a trot or a canter, as described later.

  In S602 to S606, the sub-microcomputer 104 calculates the walking cycle of the horse. First, the sub-microcomputer 104 determines whether or not the value α of the acceleration in the vertical direction has a minimum value (S602). FIG. 16B shows a flowchart of the process in S602. In S701 to S703, the sub-microcomputer 104 holds the values of the past three samples for the acceleration value α in the vertical direction. In the figure, among the values of the three samples, "left", "mid" and "right" are referred to in order from the oldest one in time. In S704, the sub-microcomputer 104 determines whether the value of "mid" among the values of the three samples is small. Strictly speaking, in S602, it is judged whether or not the value of α one sample before takes a minimum, not whether or not the current value of α is a local minimum, but there is no problem in principle. Note that, in order to determine whether the value of α has a minimum value in the processing of S701 to S704, filtering may be appropriately applied before the processing of S701 to S704 so that α exhibits a smooth time change. .

  In S602, when it is determined that α does not take the minimum value (S602: N), the sub-microcomputer 104 increments the walking cycle counter cnt (S603). If it is determined that α has a local minimum value (S602: Y), the sub-microcomputer 104 advances the process to S604. The value of the counter cnt of the walking cycle when it is determined that α takes the minimum value is the value integrated in S602 and S603 from when the α takes the previous minimum value until the current minimum value takes it. Thus, if the value of cnt is small, the current horse's gait is likely to be a short walking cycle. Therefore, when the value of cnt is smaller than the predetermined threshold (S604: Y), the accuracy n_trot of the trot is added by the value N1 (S605). After the processing of S605 and when the value of cnt is equal to or greater than the predetermined threshold (S604: N), the value of cnt is returned to 0, which is the initial value (S606).

  Subsequently, in S607 to S610, when the value β of the output obtained from the inner product arithmetic processing unit 104 is smaller than the predetermined threshold (S607: Y), the sub-microcomputer 104 adds the probability n_run of traveling by the value N2 (S608) If β is larger than the predetermined threshold (S609: Y), the gallop accuracy n_gallop is added by the value N3 (S610). Next, the process proceeds to S611.

  In S611, specifically, as shown in FIG. 16C, when one of the above-mentioned certainty is A and the upper limit is N, processing is performed so that the value A does not exceed N in S801 to S804. By the processes of S801 to S804, the values of the respective accuracies added at S605, S608, and S610 become small values when the determinations of S604, S607, and S609 are N. That is, after the horse has stopped any of trot, canter, or gallop, the corresponding probability decreases with the passage of time. After the process of S611, the process proceeds to S612. In S612 to S618, each accuracy is compared with a predetermined threshold, and if the accuracy is greater than the predetermined threshold, it is determined that the current horse's gait corresponds to the accuracy.

  FIG. 17 shows an example of an experimental result of gait determination in the present embodiment. In the figure, a graph 400 indicates time-series information of the sensor output of the acquired acceleration sensor 110. The graphs 401 and 402 respectively show the time change of the vertical vibration α and the square of the absolute value β after the process shown in FIG.

  At time 405, the magnitude of the acceleration is smaller than the threshold used in S607. Therefore, the running accuracy n_run is increasing by the process of S608 (407 in the figure). Here, the graph 403 is a plot of the interval 408 of the local minimum values of the vertical vibration. The walking cycle has become shorter after time 409 (410), and the accuracy of trot n_trot (411) is increased by the processing of S604 and S605.

  At time 412, the gait is determined to be a trot because the traveling accuracy n_run and trot accuracy r_trot exceed the threshold 413 used in S614 and S615 (414: trot). Also, since the walking cycle has become longer at time 415, the accuracy n_trot of the trot has been lowered by the determination of N in S604 and the processing of S611, but since it is still larger than the threshold 413, the gait is determined to be a trot. continuing. Next, at time 416, the absolute value β of the acceleration increases and reaches the threshold 417. Therefore, the gallop accuracy n_gallop is increased by the processing of S609 and S610. Then, since the accuracy n_trot of the trot has become smaller than the threshold 413 at time 418, the gait is determined to be contour by the processing of S614 and S615 (419). Furthermore, at time 420, since the probability of gallop n_gallop exceeds the threshold value used in S612, the gait is determined to be a gallop (421).

Third Embodiment
Next, a third embodiment will be described with reference to the drawings. Also in the present embodiment, the portable terminal 10 ′ ′ having the same configuration as the mobile phone terminal 10 of the first embodiment is used, and in the following description, the configuration equivalent to that of the first embodiment has the same reference numerals. The detailed description is omitted using. In the third embodiment, the front of a horse is determined as an example. In canter and gallop, the movement of the left and right legs of the horse is asymmetric, and there is a division between the right front and the left front depending on which of the front legs is put forward. Generally, when a horse runs on a track, the right curve is said to be advantageous in the right curve and the left curve is advantageous in the left curve, but it is the front side that is good with individual horses like human right handedness and left handedness Are said to be different. Therefore, especially in the racehorse world, the concept of being in front of the gait is also emphasized. Therefore, in the present embodiment, processing is performed to specify the current dominant foot of a target moving body (here, a horse as an example).

  In the present embodiment, as shown in FIG. 18, the portable terminal 10 ′ ′ is attached to the chest between the front legs of the horse. Therefore, in the present embodiment, the portable terminal 10 ′ ′ rotates in the direction of the arrow when the right leg comes forward, and rotates in the opposite direction when the left leg comes forward. FIGS. 19A and 19B show the relationship between the vertical movement of the horse and the rotation of the portable terminal 10 ′ ′. FIG. 19A shows the case where the horse is moving in front of the right hand, and FIG. 19B shows the case where the horse is moving in front of the left hand. The horses will advance from (1) to (8) in the figure. Further, in the figure, the solid line indicates the right leg, and the broken line indicates the left leg.

  FIG. 20 shows a part of the configuration of the sub-microcomputer 104 in the present embodiment. In the present embodiment, the sub-microcomputer 104 includes an LPF 104r, an HPF 104s, and an inner product arithmetic processing unit 104t and 104u. The sensor output of the acceleration sensor 110 is input to each of the LPF 104 r and the HPF 104 s. The outputs of the LPF 104 r and the HPF 104 s are input to the inner product calculation processing unit 104 u. The output of the LPF 104r is also input to the inner product arithmetic processing unit 104t. A series of processes performed by the LPF 104r, the HPF 104s, and the inner product calculation processing unit 104u are the same as those in the first embodiment, and thus the description thereof is omitted. Further, the sensor output of the angular velocity sensor 111 is input to the inner product calculation processing unit 104t.

As described with reference to FIG. 8, the inner product calculation processing unit 104 t outputs the angular velocity around the vertical axis in the global coordinate system of the horse here. The gravitational acceleration component vector G ′ ′ ′ and the vibration component vector S ′ ′ ′ are respectively output from the LPF 104 r and the HPF 104 s. Therefore, the magnitudes of the vibration component vectors in the vertical direction in the horse's global coordinate system are output from the inner product calculation processing unit 104 u. For convenience, it is assumed that the value of the output obtained from the inner product operation processing unit 104t is γ, and the value of the output obtained from the inner product operation processing unit 104u is δ.

  FIG. 21 is a graph showing the time change of the value of the output obtained from the inner product operation processing unit 104t as γ, and the value of the output obtained from the inner product operation processing unit 104u as δ. The phase difference between the change in vertical vibration and the change in angular velocity in the rotational direction is 180 ° out of phase according to the change in acceleration in the vertical direction and the front of the horse. Therefore, taking the product of the magnitude δ of the vibration component vector in the vertical direction at the maximum and minimum with the magnitude γ of the angular velocity around the axis in the vertical direction and determining whether the value is positive or negative, the above deviation is obtained. It can be detected.

  22A and 22B show a flowchart of the near side determination in the present embodiment. Note that this flowchart is repeatedly executed at predetermined intervals using values γ and δ obtained by the process shown in FIG. In S901, it is determined whether or not δ indicates an extreme value. Specifically, it is determined by the processing of S1001 to S1005 in FIG. 22B whether δ indicates an extreme value. The processes of S1001 to S1005 are the same as those of FIG. 16B, and thus the description thereof is omitted here. If δ does not indicate an extreme value (S901: N), the sub-microcomputer 104 calculates the product of γ and δ, and holds this product as k (S912). If δ indicates an extreme value (S901: Y), the sub-microcomputer 104 determines whether the value of k is positive or negative (S902). Here, the value of k held by the sub-microcomputer 104 is the value of k calculated by the sub-microcomputer 104 in the process one cycle before when the series of processes of S901 to S917 is one cycle. This is because, as described above, it is actually the value of k one sample before that which is the target of determination of whether or not an extreme value is indicated in the process of S902.

  If k is positive (S902: Y), it is checked whether or not the near accuracy rl indicates the likeness to the left (S903). In this embodiment, it is assumed that the accuracy in front of the left hand is high when the value of the front accuracy rl is positive, and the accuracy between the right is high when the value of the front accuracy rl is negative. As an example, the value of the previous certainty rl is set to 0 as its initial value, and fluctuates by the processing of S905 to S911. In S903, if the near accuracy rl is 0 or more, that is, it seems to be the front left, the value N4 is added to the near accuracy rl (S904). In the present embodiment, the preceding accuracy rl is limited so as not to be larger than the value N5 by the processing of S905 and S906.

  On the other hand, when k is a positive value and the near accuracy rl is a negative value, that is, the right-hand front likelihood (S902: Y, S903: N), the near accuracy rl is halved (S907). Even when k is a negative value (S902: N), as shown in FIG. 23A, processing is the same as S903 to S907 although there is a difference such as addition and subtraction, and comparison accuracy rl to -N5. (S907 to S911). After the processes of S906, S907, and S911, the process proceeds to S912. Furthermore, after the process of S912, the process proceeds to S913.

  In S913 to S917, the near accuracy rl is compared with a predetermined threshold, and if the near accuracy rl is larger than the predetermined threshold (S913: Y), it is determined that the left hand is in front (S917). In addition, if the value -rl obtained by adding minus to the previous accuracy rl is smaller than the predetermined threshold (S914: Y), it is determined that the right hand is in front (S916). Furthermore, in S913 and S914, if the value of the near certainty rl is not greater than the predetermined threshold and -rl is not less than the predetermined threshold, it is determined that the near is indeterminate (S915).

FIG. 23 shows an experimental result of the front determination in the present embodiment. The graphs 701 and 702 show the time change of the sensor output of the acceleration sensor. Also, the graphs 703 and 704 show the time change of the sensor output of the angular velocity sensor. Furthermore, 705 and 706 indicate the angular velocity γ around the vertical axis and the magnitude δ of the vertical vibration component vector obtained by the process shown in FIG. As shown in FIG. 23, the time change of the values of γ and δ is in the opposite phase in the graph of 705 and in the phase of 706. Furthermore, the graphs 707 and 708 show the time change of the near accuracy rl, respectively. In the graph 707, the near accuracy rl decreases as time passes, and in the 708 graph, the near accuracy rl increases as time passes.

  As shown in FIG. 23, at time 709 or 710, the near accuracy rl exceeds the threshold 711 or 712 in S913 or S914 (S913: Y or S914: Y), and it is determined that the right front and the left front, respectively. After that, at times 713 and 714, as can be seen from the graphs 705 and 706, the phase difference between γ and δ is not stable, and therefore the near accuracy decreases, and it is determined that the near is uncertain by the processing of S913 to S915. ing.

  By the way, for example, the circuit shown in FIG. 24 is adopted as a circuit for calculating information related to the angular velocity around the pitch axis and information related to the angular velocity around the roll axis in the sub-microcomputer 104 of the first embodiment. It can also be done. As shown in FIG. 24, the sub-microcomputer 104 includes an LPF 1004a, a band pass filter (BPF) 1004b, an inner product calculation processing unit 1004c, 1004h, 1004j, 1004p, 1004q, an outer product calculation processing unit 1004d, 1004i, a multiplication unit 1004e, 10041, 1004 n, 1004 o, sign determination units 1004 g, 1004 m, and phase shift filter 1004 k. Here, main features of the circuit shown in FIG. 24 will be described. The magnitude of the acceleration vector in the vertical direction of the user is output from the inner product calculation processing unit 1004 c. Further, the magnitude of the acceleration vector of the user in the back and forth direction is output from the inner product calculation processing unit 1004 j, and the phase shift filter 1004 k shifts the phase by a predetermined amount with respect to the output of the inner product calculation processing unit 1004 c. By multiplying the output of the arithmetic processing unit 1004 j, the sign determination unit 1004 m can determine the positive / negative of the vector K in the front-rear direction of the user and the vector H in the left-right direction.

  Therefore, the left vector L is obtained from the multiplication unit 1004n, and the front vector F is obtained from the multiplication unit 1004o. Note that depending on the initial value of the circuit shown in FIG. 24, the right vector is obtained from the multiplication unit 1004n, and the subsequent vector is obtained from the multiplication unit 1004o. Further, by obtaining the inner product of these vectors and the sensor output of the angular velocity sensor 111, from the inner product calculation processing units 1004p and 1004q, information related to the angular velocity around the pitch axis whose rotational direction is determined and the angular velocity around the roll axis Information related to each is output.

Fourth Embodiment
Next, a fourth embodiment will be described. When breeding industrial animals such as cows and horses, their health management is regarded as important. In the case of tetrapods in particular, it is required to grasp the health status of each leg. For example, when the condition of one leg is bad, the walking condition or the balance of the body in walking or running of the animal becomes bad, and the walking condition of the animal may be regarded as crawling. For example, the lameness of cattle is often caused by gonorrhea that causes sore, and in the case of a dairy cow, lameness may reduce the yield of raw milk. Also, for example, the coasting of a horse is often caused by a tendonitis in which the tendon of the forelimbs aches, and if it is a racehorse, it may lead to a situation where it can not run. Therefore, in rearing four-legged industrial animals, it is required to detect abnormalities of the legs early from the lame line and to take prompt measures.

In recent research, it has been proposed to use two evaluation criteria, regularity and symmetry, for the purpose of quantifying the horse's gait, for example. For example, when the walking state of a horse is a trot, regularity and symmetry are determined using the vertical acceleration of the horse's body. The term "trot" refers to a state in which the right front leg and the left rear leg, and the left front leg and the right rear leg make a pair and walk. The change in vertical acceleration of the front leg at this time is represented in time series as shown in FIG. In the figure, since the movement of the leg and the period of one wave are linked, the wave of each period indicates that the support leg alternates between right, left, right, and left. Therefore, as shown in FIG. 26, the correlation value of waves (waves of one and two steps before) between adjacent periods is symmetrical, and waves between periods separated by one period (one and three steps before). Wave) correlation value is defined as regularity.

In FIG. 26, the correlation calculation for calculating regularity and symmetry can not be properly executed simply by sampling the vertical acceleration at a constant interval simply because the wave period is not constant, and the following calculation method is adopted Do. In sampling of the vertical acceleration, A (m) is calculated by the following equation (11), assuming that the value before i samples is xi (i is a natural number).

  Here, N is the number of samplings. A (m) means the correlation value between the sample at the current time and the sample at a time retroactive to the past by m samples. Therefore, when a graph in which m is taken on the horizontal axis and A (m) is taken on the vertical axis, the shape starts from the mountain at m = 0 and continues to the valley, the mountain, the valley. In FIG. 27, the first peak of m = 0 is the correlation with the own sample. Further, assuming that m in the first and second peaks of the graph in FIG. 27 is s and r, respectively, A (s) and A (r) respectively correspond to the wave of one cycle before in FIG. The correlation is the correlation with the wave two cycles earlier. Here, the magnitude of the value of A (m) changes with the amplitude of the wave, so the symmetry is defined as A (s) / A (0) and the regularity as A (r) / A (0). Here, since the correlation value between a certain sample and another sample does not become higher than the correlation value in the correlation with the own sample, in principle these defined values become 1 or less.

  By the way, this calculation method has some problems. The first problem is that the amount of calculation is large. For example, assuming that N = 100, calculation of A (0) is calculated 100 times for performing one correlation operation, calculation of A (1) is calculated 99 times, etc., for a total of 5050 calculations Do. And in the gait determination of a horse, in order to repeat correlation calculation for every one step, the processing load accompanying calculation is large. The second problem is that local errors in calculating the values of regularity and symmetry become large if sampling is not performed at sufficiently short intervals with respect to the wave period. For example, in FIG. 27, it is assumed that N = 7 and the data strings x1, x2,..., X7 are 1, 0, -2, 0, 2, 0, -2, respectively. At this time, A (0) is (1 + 0 + 4 + 0 + 4 + 0 + 4) / 7 = 13/7, that is, about 1.86. On the other hand, A (4) is (2 + 0 + 4) / (7-4) = 6/3 = 2, which is a larger value than A (0).

FIG. 28 is a graph showing the results of calculating the regularity and symmetry for each step with N = 80 for the walking data of a horse. The number of acquired samples per step is about 20 samples. In the graph of FIG. 28, there is a portion where the correlation value of regularity exceeds 100%, as indicated by the area (A) surrounded by a dotted line. Also, for regularity and symmetry, it is difficult to determine which samples are normal because the samples are distributed throughout the graph. By the way, this problem can be solved by shortening the sampling interval, for example, setting the number of acquired samples in one step to 200 samples and setting a larger value, for example, 1000, as the value of N. However, if the sampling interval is thus shortened and a large value is set to N, for the sensor that detects the walking state of the horse used for sampling, the processing speed of the data collection loop is required, and the above correlation The computational complexity of the operation increases as the value of N increases. For example, in the case of N = 1000, it is estimated that about 500,000 correlation operations are performed.

So, in this embodiment, not the waveform of vertical acceleration per step but kinetic energy per step is compared. For example, assuming that the kinetic energy n steps before the current is E (n-1), the value indicating symmetry is the kinetic energy one step before (E (0)) and the kinetic energy two steps ago (E (1) The ratio of) and the value indicating regularity are obtained as a ratio of one step before and three steps before (E (2)). Here, the ratio is obtained as a value of 1 or more by using the smaller one of the kinetic energy of each step as the denominator. Further, kinetic energy, the velocity v, is given by ^mv 2/2 the mass of the object as m. In this embodiment, assuming that the velocity before α step is v _α and the velocity before β step is v _β , the ratio E (α) / E (β) of kinetic energy between α step and β step before ) is given by ^{_{(mv α 2/2) /}} (mv β 2/2) = v α 2 / v β 2.

  And since kinetic energy is the sum of squares of the area of the area surrounded by the graph of vertical acceleration and the time axis, in the present embodiment, the ratio of the square of the integral value of vertical acceleration per step is determined The ratio of kinetic energy between each step is determined. As shown in the graph of FIG. 29, the vertical acceleration at each step takes positive and negative values with respect to the time axis. Therefore, since the kinetic energy for each step is the sum of the upward kinetic energy and the downward kinetic energy, it is sufficient to find the sum of squares of the upward velocity and the downward velocity.

Here, the velocity of kinetic energy representing the walking state is obtained by integrating the acceleration. In FIG. 30A, the areas of the regions A and B indicate speeds. Since the area of the region A can be approximated to the area of the region B, kinetic energy in walking may be represented by squaring A or B. However, expressing kinetic energy using A + B as the velocity in one step in the walking state averages the velocity values more than using either A or B. Further, in the present embodiment, in evaluating the regularity and symmetry of the walking of the horse, it is determined whether the ratio of kinetic energy exceeds the threshold value. Therefore, the ratio of the velocity itself may be determined without determining the value of the velocity squared. For example, assuming that _{V 1} and _{V 2} are both positive or _negative, ^{that V} 1 2 _{/ V} ^{2 2} to determine whether 2 or _more, whether V 1 / _{V 2} is the square root of 2 or more Equivalent to determining. Therefore, in the present embodiment, as shown in FIG. 30B, the sum of squares of the speeds one step before, two steps before, and three steps before is S1, S2, and S3, respectively, and the symmetry value is S2 / S1. (S2> S1) or S1 / S2 (S1> S2), the value of regularity is defined as S3 / S1 (S3> S1) or S1 / S3 (S1> S3). In addition, each of S1, S2, and S3 which are the sum of squares of speed is a value proportional to the kinetic energy one step before, two steps before, and three steps before from the present. Also, the sum of squares of the velocity corresponds to an example of a value indicating kinetic energy.

  Next, with reference to FIG. 31, the measurement terminal 300, the jockey terminal 400 that communicates with the measurement terminal 300, the Baba apparatus 500, and the in-bare terminal 600 according to this embodiment will be described. The measurement terminal 300 includes a three-axis acceleration sensor 301, a display device 302, a communication device 303, and a sub-microcomputer 304. The measuring terminal 300 corresponds to an example of a measuring device.

  The three-axis acceleration sensor 301 is a MEMS (Micro Electro Mechanical Systems) sensor which is also used in the above embodiment. The three-axis acceleration sensor 301 measures the acceleration of the three-dimensional space accompanying the motion of the horse on which the measurement terminal 300 is mounted, and outputs it as acceleration data. Data measured by the three-axis acceleration sensor 301 is sent to the CPU 304 b via the interface 304 a. The display device 302 displays the data received from the CPU 304 b. The communication device 303 transmits the data obtained as the processing result of the CPU 304 b to the jockey terminal 400, the horse station device 500, and the horse terminal 600, for example, by wireless communication. Note that the communication method of the communication device 303 can be appropriately adopted such as wired communication or short distance wireless communication according to the use environment.

  The sub-microcomputer 304 is a general sub-microcomputer, and a main memory 304c which is a non-volatile memory and a flash memory 304d which is a volatile memory are connected to the CPU 304b. Further, the CPU 304 b transmits and receives data to and from the three-axis acceleration sensor 301, the display device 302, and the communication device 303 via the interface 304 a. The CPU 304 b receives acceleration data in a three-dimensional space from the three-axis acceleration sensor 301, and calculates the values of the regularity and symmetry described above, as well as the number of steps and gait. The processing result by the CPU 304b is stored in the main memory 304c, and then output to the external jockey terminal 400, the horse station device 500, the in-door device 600, etc. via the communication device 303 in real time or at any timing.

  As real-time output, the jockey who is riding on the horse from the measuring terminal 300 when the current walking state of the horse, that is, regularity or symmetry is regarded as abnormal by the CPU 304b as described below There is a case where a message or the like is transmitted to the jockey terminal 400 to be carried. In this case, when the jockey terminal 400 receives the message or the like, it notifies the fact that an abnormality has occurred in the horse's walking through the warning device 401 by vibration, sound, light or the like. Similarly, a jockey or trainer who is riding a horse through the warning light 501 of the Baba apparatus 500 installed in the Baba or the display device 601 of the in-bare terminal 600 that an abnormality has occurred in the walking of the horse. You may let

  In addition, when the notification in real time is not performed, the above-described message received from the measuring terminal 300 is stored as a log in the storage device 602, and the walking state of the horse at an arbitrary timing such as when the user operates the in-yard terminal 600. It can also provide information on The log is not limited to the in-yard terminal 600, but may be stored in any server that can be connected to the measurement terminal 300, the jockey terminal 400, the Baba apparatus 500, and the in-baar terminal 600.

  In the present embodiment, the application or the like stored in the main memory 304 c of the measurement terminal 300 is executed by the CPU 304 b. As shown in FIG. 32, the measurement terminal 300 functions as a gait identification unit 3001, a gait determination unit 3002, a kinetic energy value calculation unit 3003, and a kinetic energy ratio calculation unit 3004 by execution of such an application or the like. However, one or more of the gait identification unit 3001, the gait determination unit 3002, the kinetic energy value calculation unit 3003, and the kinetic energy ratio calculation unit 3004 may be a hardware circuit.

  The gait identification unit 3001 identifies the gait of the horse on which the measurement terminal 300 is mounted, using the acceleration data acquired from the three-axis acceleration sensor 301. The gait determination unit 3002 uses the gait identified by the gait identification unit 3001 to determine whether the movements of the right front leg and the left rear leg and the left front leg and the right rear leg of the horse are paired, that is, the trot It is determined whether or not it is medium. The kinetic energy value calculation unit 3003 uses the acceleration data acquired from the three-axis acceleration sensor 301 to calculate a value indicating the velocity of each period, that is, kinetic energy, for three consecutive periods in the time change of the acquired acceleration data. Do. The kinetic energy ratio calculation unit 3004 uses the velocity calculated by the kinetic energy value calculation unit 3003 to calculate the ratio of the velocity between adjacent periods in three consecutive periods and the ratio of the velocity between the periods separated by one period Do. In addition, the ratio of the velocity between these periods corresponds to an example of the ratio of kinetic energy. Details of processing of each part will be described later.

Next, processing executed by the CPU 304 b in the present embodiment will be described with reference to the flowchart in FIG. In the following description, the vertical acceleration a ₀ means the value of vertical acceleration obtained when executing the processing of the flowchart shown in FIG. 33 in the calculation process of the previous vertical acceleration. Further, the vertical acceleration a means the value of the vertical acceleration obtained when the processing of the flowchart shown in FIG. 33 is executed in the calculation processing of the current vertical acceleration.

In S1101, the CPU 304b calculates the vertical acceleration a of the horse on which the measurement terminal 300 is mounted, based on the measurement value of the three-axis acceleration sensor 301. Then, the process proceeds to S1102. In S1102, CPU304b determines whether the product of the vertical acceleration a obtained by the calculation in the vertical acceleration _{a 0} and the current S1101 obtained by the previous calculation is a negative value. By the process of S1102, it is known whether the value of the vertical acceleration has become 0 between the previous calculation and the current calculation. That is, when the product of a ₀ and a takes a negative value, it means that the vertical acceleration changes from positive to negative or from negative to positive. Also, when the product of a ₀ and a takes a positive value, it means that the positive and negative of the vertical acceleration remain positive or negative. In S1102, if the product of _{a 0} and a takes a negative value (S1102: Yes), the process advances to S1105. Further, in S1102, if the product of _{a 0} and a does not take a negative value (S1102: No), the process advances to S1103.

  In S1103, the CPU 304b adds the value a of the vertical acceleration calculated in S1101 to the integral value S of the vertical acceleration. As described above, the half cycle corresponds to the half step in the time change of the vertical acceleration. Therefore, the value of the integral value in the half cycle of the vertical acceleration can be obtained by obtaining the cumulative sum of the vertical acceleration until the vertical acceleration switches from positive to negative or negative to positive. When the process of S1103 is completed, the process proceeds to S1104.

  In S1105, the CPU 304b performs the same processing as that of the second embodiment using the vertical acceleration calculated in S1101 to determine the gait determination, gait determination, and step count measurement of the horse on which the measurement terminal 300 is mounted. Do the processing. By executing the processing of S1105, it is also possible to know whether the horse is currently in the trot. The process then proceeds to S1106.

In step S1106, the CPU 304b determines whether the vertical acceleration a calculated in step S1101 is greater than zero. If the vertical acceleration a is smaller than 0 (S1106: No), the process proceeds to S1107. In S1107, the CPU 304b stores the square of the cumulative sum S of the vertical accelerations in the square sum S1 of the speed one step before. In the present embodiment, the fact that the vertical acceleration a is switched to a value smaller than 0, that is, a negative value can be regarded as completion of the first half of the cycle in one step of the horse. Therefore, in the present embodiment, the speed of the first half of the cycle in one step of the horse is calculated as S1 by the processing of S1106 and S1107. Next, the process proceeds to S1108, where the current value of the vertical acceleration a is stored as the value of the cumulative sum S of vertical accelerations. Further process advances to S1104, the vertical acceleration a ₀ obtained by the previous calculation, the current value of the vertical acceleration a is stored. This is ready to calculate the speed for the second half of the horse's one-step cycle.

  In the present embodiment, the processing of S1101 to S1104 is repeatedly performed at predetermined time intervals. Therefore, after the speed for the first half of the horse in one step is calculated by S1106 and S1107, and the processing of S1108 and S1104 is completed, the CPU 304b executes the processing of S1101 again. Then, by repeatedly executing S1102 to S1104, the cumulative sum S of the vertical accelerations in the second half period in one step of the horse is calculated. When the second half of the horse's walking cycle is completed, the value of the vertical acceleration a calculated in S1101 switches from negative to positive. Therefore, the process proceeds from S1102 to S1105, S1106, and S1109.

In S1109, based on the result of the gait determination in S1105, it is determined whether the current gait of the horse is a trot. Generally, in the gait of a horse, in the gait other than trot, it is considered to be normal gait that the above-mentioned symmetry is asymmetric. Therefore, in the present embodiment, when it is determined in S1109 that the current gait of the horse is a trot, the CPU 304b executes the processing of S1111 and subsequent steps shown in FIG. If it is determined that the current gait is not a trot (S1109: No), the process proceeds to S1110. In S1110, the CPU 304b sets the value of the sum of squares S1, S2, and S3 of the speeds one to three steps before to 0.

  If it is determined that the current gait is a trot (S1109: Yes), the process proceeds to S1111. In S1111, the square of the cumulative sum S of the vertical accelerations is added to the square sum S1 of the speed one step before. As described above, since the speed for the first half of the horse's one-half cycle is stored in S1 in S1107, S1 is the horse's speed by adding the speed for the second half of the cycle to S1 in S1111. The speed for one cycle in one step is stored. Next, the process proceeds to S1112.

  In the present embodiment, the regularity and symmetry of the walking state of the horse are calculated using the ratio of S1 to S3 which is the sum of squares of the respective speeds one to three steps before. As the values of S2 and S3 may be stored as 0 in S1110, if S2 and S3 are 0, the kinetic energy ratio can not be determined correctly. So, in this embodiment, after confirming that S2 and S3 are not 0 in S1112 and S1113, the value of regularity and the value of symmetry are calculated.

  In S1112, the CPU 304b determines whether the value of the sum of squares S2 of the speed two steps before the current time is zero. If the value of the square sum S2 of speeds is not 0 (S1112: No), the CPU 304b advances the process to S1114. If the value of the sum of squares of speeds S2 is 0 (S1112: Yes), the CPU 304b advances the process to S1113. In S1114, the CPU 304b determines whether the value of the sum of squares S1 of speeds is larger than the value of S2. If the value of the sum of squares of speeds S1 is larger than the value of S2 (S1114: Yes), the CPU 304b advances the process to S1115. If the value of the sum of squares S1 of the speeds is not larger than the value of S2 (S1114: No), the CPU 304b advances the process to S1116. In S1115 and S1116, the CPU 304b determines the ratio of S1 and S2 using the larger value of the sum of squares S1 and S2 of the velocity as the denominator according to the determination result in S1114, and divides the determined ratio by 100 for symmetry. Calculated as sex. After the processing of S1115 or S1116, the CPU 304b advances the processing to S1113.

  In S1113, the CPU 304b determines whether the value of the sum of squares S3 of the speed three steps before the current time is zero. If the value of the sum of squares S3 of speeds is not 0 (S1113: No), the CPU 304b advances the process to S1117. If the value of the sum of squares S3 of the speeds is 0 (S1113: Yes), the CPU 304b advances the process to S1120. In S1117, the CPU 304b determines whether the value of the sum of squares S1 of the velocities is larger than the value of S3. If the value of the sum of squares of speeds S1 is larger than the value of S3 (S1117: Yes), the CPU 304b advances the process to S1118. If the value of the sum of squares S1 of the speeds is not larger than the value of S3 (S1117: No), the CPU 304b advances the process to S1119. In S1118 and S1119, the CPU 304b determines the ratio of S1 and S3 using the larger value of the sum of squares S1 and S3 of the velocity as the denominator according to the determination result of S1117, and multiplies the determined ratio by 100 as a rule Calculated as sex. After the processing of S1118 or S1119, the CPU 304b advances the processing to S1120. In S1115, S1116, S1118, and S1119, although the ratio of the sum of squares of velocity (ratio of kinetic energy) is multiplied by 100, the reason for multiplying 100 is to simply make it easy to perform integer arithmetic, and There is no need to multiply.

The CPU 304b stores the current value of S2 in S3 in S1120, and then stores the current value of S1 in S2 in S1121. As a result, the speed one or two steps before the current is made the sum of squares of the speed two or three steps before, and the sum of the squares of the speed one step before calculated newly by the processing of S1101 to S1111 is S1. It can be stored. After S1121, the CPU 304b advances the process to S1108. The processes of S1108 and subsequent S1104 are the same as described above, and thus the description thereof is omitted. As described above, the processing of S1112 to S1119 may be performed once for each sampling of the 3-axis acceleration sensor 301, and there is no loop processing, so that the correlation calculation can be performed with a low processing load compared to the conventional correlation calculation. it can.

  In FIG. 34, based on the results obtained by executing the process of the above-described flowchart, four horses A and B having normal walking and four horses C and D of the herd who are regarded as abnormal by the process of this embodiment The figure which plotted the regularity and symmetry in the gait of is. As shown in the figure, in the case of horses A and B with normal walking, both regularity and symmetry are often plotted before and after the value 100. On the other hand, in the case of the horses C and D, it is understood that there is a tendency to be plotted in a region where the symmetry is particularly bad. The walking data of horse B in FIG. 34 is also used in the graph of regularity and symmetry shown in FIG. As compared with the graph of FIG. 28 and the graph of horse B of FIG. 34, the graph of FIG. 34 obtained in the present embodiment is plotted closer to the value 100 at which both the symmetry and regularity can be regarded as better. Therefore, in the present embodiment, it can be said that it is possible to determine whether the horse's walking is normal more accurately than in the past based on the regularity and symmetry of the horse's walking obtained by the above processing. .

  Next, in the present embodiment, based on the regularity and symmetry determined by the process of FIG. 31, the process of detecting whether or not the current walking state of the horse is above will be described with reference to FIG. Do. In the present embodiment, in order to warn of the result of detection of coasting, an abnormality is detected based on the values of regularity and symmetry obtained by executing the flowchart shown in FIG. As a method of reducing CPU processing load, it is possible to check whether the values of symmetry and regularity frequently exist outside the area considered to be normal. In the following, according to the present embodiment, a score indicating regularity and symmetry in the walking of a horse is calculated, and based on the score, it is determined whether the walking of the horse is lame or not, and a process of giving a warning when lame Will be explained.

  In this embodiment, after the values of regularity and symmetry are calculated by executing the flowchart shown in FIG. 33, the CPU 304b can start the processing of the flowchart shown in FIG. 35 at an arbitrary timing. Further, the CPU 304 b repeatedly executes the processing of this flowchart at predetermined time intervals determined based on the required processing capability, the accuracy of the walking determination, and the like. In the present embodiment, the larger the regularity or symmetry value, the higher the possibility that it can be regarded as coasting. Therefore, for the values of regularity and symmetry, a threshold as an upper limit value that can be regarded as normal walking is determined in advance and stored in the memory 304 c.

  Further, in the flowchart of FIG. 35, a score (“Scr” in the figure) is used as an index indicating whether or not walking is crawling. In the present embodiment, the initial value of the score is 0. Further, in the present embodiment, each value of Z1, Z2, and Z3 used in the following description is set in advance. FIG. 36 shows an example of a graph plotting each score when the process of FIG. 35 is repeatedly executed. The score plotted in FIG. 36 is the value of the score obtained each time the flowchart of FIG. 35 ends. Note that, as shown in FIG. 36, Z1 is a score increase value, Z2 is a threshold larger than Z3, and it is regarded as an insult when the score takes a value from Z3 to Z2 Period)).

  In S1201, the CPU 304b determines whether the calculated regularity and symmetry values are both smaller than the above-described threshold value acquired from the memory 304c. If the values of regularity and symmetry are both smaller than the threshold (S1201: Yes), the CPU 304b advances the process to S1202. If neither the regularity nor the symmetry value is smaller than the threshold (S1201: No), the CPU 304b advances the process to S1204.

  At S1202, the CPU 304b determines whether the score is greater than zero. If the score is greater than 0 (S1202: Yes), the CPU 304b advances the process to S1203. If the score is not greater than 0 (S1202: No), the CPU 304b advances the process to S1207. In S1203, the CPU 304b subtracts 1 from the score. After the process of S1203, the CPU 304b advances the process to S1207.

  In S1204, the CPU 304b adds Z1 to the score. After the processing of step S 1204, the CPU 304 b advances the processing to step S 1205. In S1205, the CPU 304b determines whether the score is greater than Z2. If the score is greater than Z2 (S1205: Yes), the CPU 304b advances the process to S1206. If the score is not larger than Z2 (S1205: No), the CPU 304b advances the process to S1207. In S1206, the CPU 304b stores Z2 in the score. After the process of S1206, the CPU 304b advances the process to S1207.

  In S1207, the CPU 304b determines whether the score is larger than Z3. If the score is greater than Z3 (S1207: Yes), the CPU 304b advances the process to S1208. If the score is not larger than Z3 (S1207: No), the CPU 304b ends the processing of this flowchart. In S1208, the CPU 304b executes processing to warn that the current horse's walking has been determined to be a coasting. As the warning process, for example, the CPU 304 b transmits a warning message to the display device 302 via the interface 304 a. In addition, the CPU 304b transmits an instruction to cause the warning device 401 to execute a warning to the jockey terminal 400, transmits an instruction to turn on the warning light 501 to the Baba device 500, or displays the display device on the in-bare terminal 600. An instruction to display a warning may be sent to 601. After the process of S1208, the CPU 304b ends the process of this flowchart.

  As also shown in the graph of FIG. 36, in the present embodiment, the processing of S1205 and S1206 is performed, so that the score does not increase beyond Z2 even if it continues to increase. Then, the processing of S1202 and S1203 is executed, and the warning of the coasting is stopped at the stage where the score is lowered and becomes smaller than the threshold value Z3. At this time, the CPU 304 b transmits an instruction to cancel the warning to the display device 302 or the like.

  In the above process, as shown in FIG. 37A, in the normal region set when determining the calculated regularity and symmetry values, the regularity and symmetry values become a predetermined size. Is an area that is acceptable. However, as shown in FIG. 37B, for example, an abnormal area may be set instead of the normal area, and the configuration may be configured to warn of coasting when the values of regularity and symmetry are within the abnormal area. In this case, the process of S1201 described above is changed to a process of determining whether or not the values of regularity and symmetry are within the range of the values of the abnormal region set respectively. Furthermore, if the values of regularity and symmetry are within the range, the process proceeds to S1204, and if the values of regularity and symmetry are not within the range, the process proceeds to S1202.

  Furthermore, as an alternative configuration example, it is also possible to move the normal region set as shown in FIG. 37A to another range. In this case, the process of S1201 may be changed to a process of determining whether the values of regularity and symmetry are within the range of the values of the normal region. By changing the configuration in this manner, it is possible to adjust the judgment criteria of coasting according to the walking characteristics of each horse.

Fifth Embodiment
Next, a fifth embodiment will be described. In addition, about the component similar to 4th Embodiment, the same code | symbol is attached | subjected and detailed description is abbreviate | omitted. In the present embodiment, a ring buffer (not shown) is formed in the flash memory 304 d to store the values of regularity and the value of symmetry calculated from the previous N steps before to the present. There is also a belief that the asymmetry of walking is not unusual as racehorses, in particular the legs of the thoroughbred, are originally malformed. According to this idea, even in a horse whose walking is normal but asymmetric walking, the values of regularity and symmetry calculated by the above processing do not always fall within the normal region. In addition, even when adopting a configuration in which walking is determined based on whether or not the values of regularity and symmetry exist in the abnormal region as shown in FIG. 37B, the regularity and symmetry calculated according to the condition of malformation of individual horse legs The value of may be different and not a general criterion.

  So, in this embodiment, after grasping the tendency of regularity and symmetry unique to each horse, a warning is given when the tendency of regularity and symmetry changes. In the present embodiment, the index value is calculated using the values of regularity and symmetry calculated in the past, for the horse which is the target of the walking judgment, and, for example, the average of the values of regularity and symmetry from the present to 100 steps ago Compare the value with the index value. And when the difference between these values falls within the threshold, walking can be regarded as normal.

  FIG. 38 shows a flowchart executed in the present embodiment. The process of this flowchart is started by, for example, the CPU 304 b when the measurement terminal 300 is attached to a horse to be determined and the measurement terminal 300 is powered on. In S1301, the CPU 304b sets the cumulative sum Tr of the regularity values, the cumulative sum Ts of the symmetry values, the number n of additions to the Tr and Ts to 0, respectively. Furthermore, the CPU 304 b sets a value (for example, “−1”) indicating an uninitialized state that has not been initialized yet, to the counter (Index) of the above-mentioned ring buffer in the flash memory 304 d. In the present embodiment, a negative value is used as the value indicating the uninitialized state, but another value may be used as long as it can distinguish the uninitialized state. The process of S1301 may be performed manually when the measurement terminal 300 is attached instead of being executed by the CPU 304b. Alternatively, for example, when the battery (not shown) of the measuring terminal 300 has enough capacity to continue mounting the measuring terminal 300 for a few weeks, for example, for several weeks, the clock of the CPU 304 b, for example. It may be executed at a predetermined timing using an internal timer or the like using a frequency. After the process of S1301, the CPU 304b advances the process to S1100.

  In S1100, the CPU 304b executes the processing of the flowchart shown in FIG. 33 to calculate the current gait regularity and symmetry value. When the values of regularity and symmetry are calculated, the CPU 304b advances the process to step S1302. In S1302, it is determined whether the values of regularity and symmetry have been calculated and updated by S1115 or S1116, S1118 or S1119, S1120, S1121 in the process of S1100. If the values of regularity and symmetry have been calculated and updated (S1302: Yes), the CPU 304b advances the process to S1305. If the values of regularity and symmetry have not been calculated and updated (S1302: No), the CPU 304b advances the process to S1303.

  In S1305, the CPU 304b adds the value of regularity and the value of symmetry calculated in the process of S1100 to Tr and Ts, respectively, and adds 1 to the value of n. Then, the process proceeds to S1306. In S1306, the CPU 304b determines whether the value of the counter Index of the ring buffer of the flash memory 304d is negative. If the value of Index is negative (S1306: Yes), the CPU 304b advances the process to S1307. If the value of Index is not negative (S1306: No), the CPU 304b advances the process to S1312.

  In S1307, 0 is stored in the value i used as a counter of the loop processing of S1308 to S1310. The process then proceeds to S1308. In this embodiment, a ring buffer capable of storing the past N regularity values and symmetry values is provided, and the array of ring buffers for storing regularity values is Dr, the symmetry value Let Ds be the array of ring buffers that store Further, the numbers of the arrays Dr and Ds are 0 to N-1, and Dr [i] and Ds [i] mean data stored in the array whose number is i. By the loop processing of S1308 to S1310, the values of the regularity and the value of symmetry calculated at S1100 are stored in the arrays Dr and Ds of numbers 0 to N-1, respectively. When the loop processing of S1308 to S1310 is completed (S1310: No), the CPU 304b advances the processing to S1311.

In S1311, the CPU 304b stores 0 as the value of Index, and calculates the sum Sr, Ss of data stored in each array Dr, Ds of numbers 0 to N-1. Since the regularity value and symmetry value calculated in S1100 are stored in the arrays Dr and Ds of numbers 0 to N-1 by the loop process of S1308 to S1310, S1 and Ss are calculated in S1100. The regularity value and the symmetry value multiplied by N are stored respectively. In this way, it is possible to initialize an Infinite Impulse Response (IIR) filter that calculates a moving average. After the process of S1311, the CPU 304b advances the process to S1303.

  In S1312, the CPU 304b adds a value obtained by subtracting the value stored in the array Ds whose number is the value of Index from the value of symmetry calculated in S1100 to Ss. Also, the CPU 304 b adds a value obtained by subtracting the value stored in the array Dr whose number is the value of Index from the value of the regularity calculated in S1100 to Sr. Then, the CPU 304b advances the process to S1313, stores the value of symmetry calculated in S1100 in the array Ds whose number is the value of Index, and calculates S100 in the array Dr whose number is the value of Index Stores regularity values. Next, the CPU 304 b advances the process to step S1314.

  In S1314, the CPU 304b adds 1 to the value of Index. Then, the CPU 304b advances the process to step S1315 and determines whether the value of the index is N or more. If the value of the index is N or more (S1315: Yes), the CPU 304b advances the process to S1316. If the value of Index is not N or more (S1315: No), the CPU 304b advances the process to S1317. In S1316, the CPU 304b returns the value of Index to 0. Then, the CPU 304 b advances the process to step S1317. In S1317, the CPU 304b determines whether the number of additions to Tr and Ts is N or more. If n is equal to or greater than N (S1317: Yes), the CPU 304b advances the process to S1318. If n is not equal to or greater than N (S1317: No), the CPU 304b advances the process to S1303.

  In this embodiment, the moving average calculation process is initialized by the processes of S1307 to S1311, and the values of regularity and symmetry initially calculated at S1100 are stored in the arrays Dr and Ds as values for the past N times. Be done. Therefore, the sums Sr and Ss of data stored in the arrays Dr and Ds are N until the number n of additions of the cumulative sum Tr of regularity and the cumulative sum Ts of symmetry is 0 in S1301 and then N. The values of regularity and symmetry calculated in S1100 are not correctly reflected. Therefore, in S1317, when n becomes N or more, the processing of the CPU 304b is controlled so that the processing of determination of necessity of warning and execution of warning in S1318 to S1320 is executed.

Further, in the present embodiment, the measurement terminal 300 is repeatedly attached to and removed from the horse which is the target of the walking determination every day. Therefore, for example, when training or the like on a day of a horse wearing the measurement terminal 300 is finished, the measurement terminal 300 is removed from the horse. When the measurement terminal 300 is removed from the horse, the CPU 304b advances the process from S1303 to S1304 as shown in FIG. In the present embodiment, using the average value Tr / n and Ts / n of the values of regularity and the values of symmetry calculated from the installation to the removal of the measurement terminal 300, the previous day of the horse equipped with the measurement terminal 300 A value Cr indicating the tendency of regularity up to and a value Cs indicating the tendency of symmetry on the previous day are calculated. In S1303, the CPU 304b may execute a process of determining whether a predetermined time has elapsed using a timer (not shown). In this case, the values of Cs and Cr are a value indicating the tendency of symmetry until the previous wearing and a value indicating the tendency of regularity, respectively.

In S1304, the CPU 304b stores the value calculated by 0.9 × Cr + 0.1 × Tr / n in Cr as a new value indicating the tendency of regularity up to the previous day. Similarly, the CPU 304 b stores the value calculated by 0.9 × Cs + 0.1 × Ts / n in Cs as a new value indicating the tendency of symmetry until the previous day. By the processing of S1304, the average value of the regularity value and the symmetry value as of today, and the value showing the tendency of regularity and symmetry up to the previous day can be added with appropriate weighting. The weighting in the calculation of S1304 can be changed as appropriate. Further, the calculation method in S1304 is calculation in the case of using a so-called general first-order delay FIR (Finite Impulse Response) filter. However, instead of this calculation method, it is possible to obtain a value indicating the tendency of regularity and symmetry until the previous day, using a known method such as obtaining a moving average for the past several days.

  Furthermore, in the present embodiment, average values Ss / N and Sr / N over the past N times are calculated for the values of regularity and symmetry calculated in S1100, and regularity until the previous day calculated in S1304. And contrast values Cs, Cr indicating tendency of symmetry. As a comparison method, as shown in FIG. 39, with the values of regularity and the values of symmetry as the vertical axis and the horizontal axis, respectively, points indicated by the average values of the past N regularities and the values of the symmetry, The distance between the points indicated by the values indicating the tendency of regularity and symmetry until the previous day is compared with a predetermined threshold.

At S1318 in FIG. 38, the CPU 304b uses the following equation (12) to indicate the point indicated by the average value of the past N regularity and symmetry values, and the value indicating the tendency of the regularity and symmetry up to the previous day Calculate the value R which is the square of the distance between the point indicated by.
R = (Ss / N-Cs) ² + (Sr / N-Cr) ² (12)
The distance is strictly the square root of the value R. However, considering that the distance is compared with the threshold value in S1319, the processing load of the CPU 304b can be reduced by calculating without taking the square root. Calculate the value of When the process of S1318 ends, the CPU 304b advances the process to S1319.

  In S1319, the CPU 304b determines whether R calculated in S1318 is larger than a predetermined threshold. As can be seen from FIG. 39, as the value of R increases, the degree of regularity and symmetry from the present to the past N times deviates from the tendency of regularity and symmetry until the previous day. If it is determined in S1319 that R is greater than the predetermined threshold (S1319: Yes), the regularity and symmetry from the present to the past N times are regarded as abnormal, and the CPU 304b advances the process to S1320. If it is determined that R is not larger than the predetermined threshold (S1319: No), the regularity and symmetry from the present to the past N times are regarded as normal, and the CPU 304b advances the process to S1303.

In S1320, the CPU 304b warns that the present horse's walking is considered to be lameness. The warning process is the same as that of the fourth embodiment, and thus the detailed description is omitted. After performing the warning process, the CPU 304b advances the process to step S1303. In S1319, instead of determining R calculated by equation (12) using a predetermined threshold value, R1 = (Ss / N-Cs) ² and R2 = (Sr / N-Cr) ² , R1 and R2 may be compared with the respective threshold values to determine whether both R1 and R2 are greater than the respective threshold values. In addition, regularity and symmetry may be weighted by multiplying each term of equation (12) by an appropriate constant.

Sixth Embodiment
Next, a sixth embodiment will be described. In addition, the same code | symbol is attached | subjected to the component similar to said embodiment, and detailed description is abbreviate | omitted. Conventionally, for example, when the acceleration during movement is measured for a racehorse, a device equipped with a GPS sensor is mounted on the racehorse, and the acceleration is calculated based on GPS data transmitted from the device. Generally, the GPS sensor outputs GPS data including current position information and time information at regular intervals of time. And when measuring GPS data at fixed time intervals, the condition of the horse at the time of measurement, for example, expansion and contraction of the horse at the time of gallop, etc. may differ for each measurement timing. For this reason, since it is not possible to continue measuring GPS data at the timing when the positional relationship of each leg of the horse becomes the same, the variation of the positional relationship of each leg of the horse at each measurement timing is an increase or decrease of the calculated acceleration. It is reflected.

  Another possible method is to increase the measurement accuracy of acceleration by collecting more GPS data by shortening the output interval of GPS data by the GPS sensor and calculating the average of the accelerations calculated for each predetermined unit time. However, in order to increase the power consumption of the GPS sensor, this method is not realistic when it is considered to be mounted on the measurement terminal or the like in the above embodiment. Therefore, in the present embodiment, a method will be described in which a GPS device is mounted on the measurement terminal used in the above embodiment, and the acceleration is accurately calculated while suppressing the power consumption of the GPS device.

  FIG. 40 shows a measurement terminal 700, a jockey terminal 400 that communicates with the measurement terminal 700, a Baba apparatus 500, and a Baba terminal 600 in the present embodiment. The measurement terminal 700 includes a three-axis acceleration sensor 301, a display device 302, a communication device 303, a sub-microcomputer 304, a GPS sensor 305, and a timer 306. The measuring terminal 700 corresponds to an example of a measuring device. The sub-microcomputer 304 has an interface 304a, a CPU 304b, a main memory 304c, and a flash memory 304d. The GPS sensor 305 is known, and transmits position information and time information on the current position of the measurement terminal 700 to the CPU 304 b via the interface 304 a. The CPU 304 b stores the position information and time information received from the GPS sensor 305 in the flash memory 304 d. The timer 306 clocks the current time, and data related to the current time is acquired by the CPU 304 b via the interface 304 a.

  In the present embodiment, an application or the like stored in the main memory 304 c of the measurement terminal 700 is executed by the CPU 304 b. By executing such an application, as shown in FIG. 41, the measurement terminal 700 functions as a gait identifying unit 7001, a gait judging unit 7002, a position information generating unit 7003, and an acceleration calculating unit 7004. However, one or more of the gait identification unit 7001, the gait determination unit 7002, the position information generation unit 7003, and the acceleration calculation unit 7004 may be a hardware circuit.

The gait identification unit 7001 identifies the gait of the horse on which the measurement terminal 700 is mounted, using the acceleration data acquired from the three-axis acceleration sensor 301. The gait determination unit 7002 determines whether or not the predetermined leg of the horse, in this embodiment, the right forefoot has landed, using the gait identified by the gait identification unit 7001. The position information generation unit 7003 acquires, from the GPS sensor 305, position information indicating the position of the predetermined leg of the horse on which the measurement terminal 700 is mounted, which is the position of the horse when it is determined that the right forefoot lands in this embodiment. It generates using position information and time information which data shows. The acceleration calculation unit 7004 uses the generated position information and the time information indicated by the GPS data to determine again that the predetermined leg of the horse on which the measurement terminal 700 is mounted, in this embodiment, the right forefoot has arrived, and then again. An average value of acceleration in movement of the horse until it is determined that the predetermined leg of the horse, in this embodiment, the right forefoot has landed, is calculated. Details of processing of each part will be described later.

  Next, the process of calculating the acceleration in the present embodiment will be described with reference to the flowchart of FIG. The processing of this flowchart is started by, for example, the CPU 304 b when the measurement terminal 700 is attached to a horse to be measured and the power of the measurement terminal 700 is turned on. Also, the CPU 304 b repeatedly executes the processing of this flowchart at predetermined time intervals.

  In S1401, the CPU 304b performs processing similar to that of the second embodiment by using the output data of the three-axis acceleration sensor 301, thereby determining the walking and walking of the horse on which the measuring terminal 700 is mounted. Perform each process of volume judgment and step count measurement. The CPU 304 b also acquires data of the current time from the timer 306. The time indicated by the data of the current time acquired here is Tp. Thus, it is possible to detect, for example, the timing at which the right forefoot lands on the legs of the horse. In the present embodiment, the acceleration is calculated based on the timing when the right front leg of the horse on which the measurement terminal 700 is mounted lands. The following acceleration can be calculated based on the timing at which another leg lands instead of the right forefoot. After the process of S1401, the CPU 304b advances the process to S1402.

  In S1402, the CPU 304b determines whether the processing result in S1401 indicates that the right forefoot of the horse has landed. If the processing result of S1401 indicates that the right forefoot of the horse has landed (S1402: Yes), the CPU 304b advances the processing to S1403. If the processing result of S1401 does not indicate that the right forefoot of the horse has landed (S1402: No), the CPU 304b ends the processing of this flowchart.

  In S1403, the CPU 304b reads GPS data acquired from the GPS sensor 305 from the flash memory 304d before and after the time Tp for which the processing result in S1401 has been calculated. In this embodiment, for example, the latitude indicated by GPS data acquired at time Ta immediately before time Tp at which the processing result of S1401 is calculated is Xa and longitude is Ya, and the time immediately after time Tp at which the processing result of S1401 is calculated. The latitude indicated by the GPS data acquired in Tb is Xb, and the longitude is Yb. Thereby, the position of measurement terminal 700 before and behind the time when the right forefoot of the horse equipped with measurement terminal 700 landed is known. After the process of S1403, the CPU 304b advances the process to S1404.

In S1404, the CPU 304b uses the following equation (13) to calculate the corrected latitude Xp based on the latitude before and after the time Tp at which the right forefoot of the horse wearing the measurement terminal 700 lands.
Xp = Xa + (Xb-Xa) x (Tp-Ta) / (Tb-Ta) (13)
Here, Tp-Ta indicates an elapsed time (for example, in milliseconds) from when the latitude Xa and the longitude Ya of the GPS data are acquired to when the processing result is calculated in S1401. Further, Tb-Ta indicates an elapsed time (for example, in milliseconds) from when the latitude Xa and the longitude Ya of the GPS data are acquired to when the latitude Xb and the longitude Yb of the GPS data are acquired. In the present embodiment, Xp calculated by equation (13) is regarded as the latitude at time Tp. After the process of S1404, the CPU 304b advances the process to S1405.

In S1405, the CPU 304b calculates the corrected longitude Yp based on the longitudes before and after the time Tp when the right forefoot of the horse wearing the measurement terminal 700 lands, using the following equation (14).
Yp = Ya + (Yb-Ya) × (Tp-Ta) / (Tb-Ta) (14)
In the present embodiment, Yp calculated by equation (14) is regarded as the longitude at time Tp. S
After the processing of 1405, the CPU 304b advances the processing to S1406.

In S1406, the CPU 304b calculates the moving distance per unit time (for example, 1 second), that is, the speed Vp in the time Tp at which the processing result of S1401 is calculated, using the following Expression (15).
Vp = Tu × {(Xp−Xo) ² + (Yp−Yo) ² } ^1/2 / (Tp−To) (15)
Here, Tu is a constant determined based on the relationship between the unit time employed when calculating Vp and the time information of the GPS data and the unit time indicated by the time information of the timer 306. For example, if Vp is determined as the speed per second, and Ta, Tb, To, and Tp are calculated as time in milliseconds in Equations (13) and (14), then Tu = 1000. Xo and Yo mean Xp and Yp calculated by S1404 and S1405, respectively, in the previous processing. Moreover, To means the time when the process result of S1401 was calculated in the previous process. The latitude and longitude (i.e., Xo, Yo, Xp, Yp) calculated in the previous and current processes and the time (i.e., To, Tp) at which the processing result in S1401 is calculated are stored in the flash memory 304d. After the process of S1406, the CPU 304b advances the process to S1407.

In S1407, the CPU 304b uses the following equation (16) to determine the elapsed time from the time To at which the processing result of S1401 was calculated in the previous processing to the time Tp at which the processing result in S1401 is calculated in the current processing. The acceleration Ap is calculated.
Ap = (Vp-Vo) / (Tp-To) (16)
The CPU 304 b stores the calculated value of the acceleration Ap in the flash memory 304 d. When the process of S1407 ends, the CPU 304b ends the process of this flowchart.

43A and 43B, measurement data when acceleration is calculated based on GPS data acquired at constant time intervals by the above-described conventional method, and processing of the above-described flowchart in this embodiment are executed to execute acceleration. The measured data in the case of calculation are shown respectively. In the graphs shown in FIGS. 43A and 43B, the horizontal axis is time (s), and the vertical axis is acceleration (m / s ² ). In the graphs shown in FIGS. 43A and 43B, the length of time on the horizontal axis is the same.

  As shown in FIG. 43A, although the acceleration data itself can be calculated by the conventional measurement method, as described above, the walking (running) state of the horse at each calculation time, in particular, the positional relationship between the legs varies. Can not grasp the change tendency of On the other hand, according to the measurement method in the present embodiment, the acceleration is continuously calculated when the walking (running) state of the horse becomes substantially the same by detecting the timing at which the right front leg of the horse lands. For example, when each leg floats in the air during gallop, the vertical acceleration is substantially zero, and when the horse steps on the right front after the right front lands, the vertical acceleration takes a maximum value. According to the present embodiment, the horse is accurate to such an extent that the increase / decrease tendency of the acceleration can be visually confirmed as shown in FIG. 43B without being affected by the acceleration when the walking (running) state of the horse is different. Acceleration can be measured.

  In the present embodiment, the data of the acceleration calculated in S1407 may be appropriately filtered and corrected according to the purpose of evaluation of the acceleration. Also, in general, when the horse runs, the timing at which the right fore landing lands occurs approximately twice a second. Therefore, assuming that the cycle of GPS data acquisition by the GPS sensor 305 is one second, for example, landing of the right forefoot occurs multiple times within one cycle of GPS data acquisition. Even in such a case, the processing of this embodiment can calculate the position of the measurement terminal at each timing when the right forefoot of the horse lands.

  The above is the description of the present embodiment, but the configuration and processing of the above-described information processing apparatus are not limited to the above embodiment, and various modifications can be made within the scope not losing the same technical concept as the present invention. Changes are possible. Hereinafter, three modified examples of the above embodiment will be described.

(Modification 1)
In the above description in FIGS. 30A and 30B, the area enclosed by the graph of vertical acceleration and the time axis is used as the speed. Therefore, in the flowchart shown in FIG. 33, in S1107 and S1111, the area in each case when the vertical acceleration is positive and negative is calculated, and the sum of them is calculated as the speed. That is, as described using FIG. 44, S1 = d ₁ ² + d ₂ ² and S 2 = d ₃ ² + d ₄ ² are calculated when the process of S1111 is completed. Here, in the above embodiment, since the attachment height of the three-axis acceleration sensor is constant, the average value of the speeds in the vertical direction can be regarded as zero. Then, when the average value of the velocity in the vertical direction is regarded as 0 every step, it can be regarded as d ₁ = −d ₂ . Therefore, when the average value of d ₁ and d ₂ is defined by the following equation (17),

The average value of d3 and d4 can be similarly defined, and S1 and S2 can be modified to obtain equations (18) and (19).


Therefore, the symmetry and regularity can be calculated by the following equations (20) and (21), respectively.

As described in the fourth embodiment, when comparing the values of regularity and symmetry with the threshold value, the ratio of the average values may not be squared as in the equations (20) and (21). Therefore, the symmetry and regularity can be calculated by the following equations (22) and (23), respectively.


Based on equations (22) and (23), in the flowchart shown in FIG. 33, instead of calculating S1 = S × S in S1107, S1 = −S is calculated, and further S1 + = S × S is calculated in S1111 Instead, it can be changed to the process of calculating S1 + = S.

(Modification 2)
Next, Modification 2 will be described. For example, if a horse has injured any leg, the positive component of the vertical acceleration described above may be different from the positive component of the vertical acceleration one step earlier because the horse covers the injured leg. It is done. According to this idea, in FIG. 45, the result of comparing d1 and d3 can be used as one criterion to determine whether the horse has injured the leg. In the above embodiment, in addition to the values of regularity and symmetry, it may be used as a third determination criterion whether the value of the ratio between d1 and d3 is within a predetermined range. In this case, as shown in FIG. 45, the value of the ratio of d1 and d3 can be set as the third evaluation axis, and a three-dimensional abnormal region can be determined by the value of regularity and the value of symmetry.

  The value of the ratio between d1 and d3 can be calculated by omitting the process of S1111 in the flowchart of FIG. 33 and executing the processes of S1112 to S1116. Therefore, the processing of the flowchart of FIG. 33 is first performed to calculate the values of regularity and symmetry, and the processing of S1111, S1113 to S1120 is omitted in the flowchart of FIG. 33 to calculate the ratio of d1 to d3. Calculate the value of Then, when the flowchart of FIG. 35 is executed, it is determined in FIG. 45 whether or not the ratio of d1 and d3 is both small in addition to the values of regularity and symmetry in S1201. It can be determined whether or not it is within the indicated abnormal area, and a warning can be issued.

(Modification 3)
Next, Modification 3 will be described. In the above description, the GPS sensor 305 of the measurement terminal 700 generates GPS data at fixed time intervals. In this modification, the processing of the CPU 304b when calculating the acceleration in the movement of the horse wearing the measurement terminal 700 when the generation timing of the GPS data by the GPS sensor 305 can be controlled will be described. In this modification, it is assumed that the CPU 304b can instruct the GPS sensor 305 to generate GPS data at an arbitrary timing and acquire GPS data generated. The configuration of the measurement terminal of this modification is the same as that of the sixth embodiment except that the timer 306 is deleted from the measurement terminal 700, so the same components as in the sixth embodiment are the same. The reference numerals and symbols are used, and the illustration and the detailed description are omitted.

The process executed by the CPU 304 b in the present modification will be described with reference to the flowchart in FIG. In S1501, the CPU 304b performs the same processing as the second embodiment by using the output data of the three-axis acceleration sensor to determine the walking of the horse on which the measuring terminal 700 is mounted, gait determination, Perform each step measurement process. Next, the process proceeds to S1502. The processing of S1502 is the same as S1402 in the flowchart of FIG.

  In S1503, the CPU 304b acquires current GPS data from the GPS sensor 305. Next, the process proceeds to S1504. In S1504, the CPU 304b calculates the difference between the latitude of the GPS data acquired in S1503 in the previous processing and the latitude of the GPS data acquired in S1503 in the current processing. Next, the process proceeds to S1505. In S1505, the CPU 304b calculates the difference between the longitude of the GPS data acquired in S1503 in the previous process and the longitude of the GPS data acquired in S1503 in the current process. The process then proceeds to S1506.

  In S1506, the CPU 304b calculates the elapsed time from the landing time of the right forefoot in the previous processing to the arrival time of the right forefoot in the current processing from the time information of the GPS data acquired in S1503 in the previous and current processing. Furthermore, the CPU 304 b calculates the speed using the calculated elapsed time and the difference between the latitude and the longitude calculated in S 1504 and S 1505. Next, the process proceeds to S1507. In S1507, the CPU 304b uses the elapsed time calculated in S1506 and the speed calculated in S1506 to calculate an acceleration. When the process of S1507 ends, the CPU 304b ends the process of this flowchart.

  In this modification, GPS data at the time when the front of the right leg lands in S1501 and S1502 can be acquired. For this reason, the CPU 304b does not execute the correction processing of latitude and longitude based on the time information and GPS data output from the timer 306 or the calculation processing of the equations (13) to (16) as shown in the flowchart of FIG. For this reason, it is possible to calculate the horse acceleration more accurately while further reducing the processing load on the CPU 304b compared to the sixth embodiment.

<A computer readable recording medium>
A management tool for setting the information processing apparatus on a computer or other machine or apparatus (hereinafter, a computer or the like), a program for realizing an OS or the like can be recorded on a recording medium readable by the computer or the like. Then, the function can be provided by causing a computer or the like to read and execute the program of the recording medium. Here, the computer is, for example, a measurement terminal or the like.

  Here, a recording medium readable by a computer etc. is a recording medium which can store information such as data and programs electrically, magnetically, optically, mechanically or chemically and read from a computer etc. Say Among such recording media, those removable from the computer etc. include, for example, memory such as flexible disk, magneto-optical disk, CD-ROM, CD-R / W, DVD, Blu-ray disk, DAT, 8 mm tape, flash memory etc. There is a card etc. Further, as a recording medium fixed to a computer or the like, there is a hard disk, a ROM or the like.

  Further, the following appendices will be disclosed regarding the above embodiment.

(Supplementary Note 1)
To the measuring device attached to the tetrapod animal,
Obtaining acceleration data of a three-dimensional space from an acceleration sensor attached to the four-legged animal;
Causing the gait of the tetrapod to be identified using the acquired acceleration data;
Using the specified gait of the four-legged animal, it is determined whether or not the movement of the right fore leg and left back leg, and the left fore leg and right back leg of the four-legged animal are paired, respectively.
When it is determined that the movement of the right front leg and the left rear leg, and the left front leg and the right rear leg of the four-legged animal are paired, the temporal change of the acquired acceleration data using the acquired acceleration data Calculate the value indicating the kinetic energy for each cycle for three consecutive cycles in
Using the value indicating the calculated kinetic energy, calculate the ratio of kinetic energy between adjacent periods in the three consecutive periods and the ratio of kinetic energy between periods separated by one period.
Measurement program characterized by

(Supplementary Note 2)
In addition to the measuring device
When the ratio of kinetic energy between adjacent cycles and the ratio of kinetic energy between cycles separated by one cycle are smaller than a predetermined threshold, it is determined that the walking of the four-legged animal is abnormal. The measurement program according to Appendix 1.

(Supplementary Note 3)
In addition to the measuring device
An index value is calculated using the cumulative sum of the ratio of kinetic energy between the adjacent cycles in the past and the cumulative sum of the ratio of kinetic energy between the cycles separated by one cycle to calculate the kinetic energy of the current adjacent cycle If the cumulative sum of the ratio and the cumulative sum of the ratio of the kinetic energy between the cycles separated by one cycle deviate from the index value by a predetermined amount, it is determined that the walking of the tetrapod animal is abnormal. The measurement program according to supplementary note 1, characterized in that

(Supplementary Note 4)
To the measuring device attached to the tetrapod animal,
Obtaining acceleration data of a three-dimensional space from an acceleration sensor attached to the four-legged animal;
Causing the gait of the tetrapod to be identified using the acquired acceleration data;
It is determined whether or not a predetermined leg of the four-legged animal has landed using the specified gait of the four-legged animal,
Position information and time information indicated by GPS data acquired from a GPS sensor attached to the four-legged animal, position information indicating the position of the four-legged animal when it is determined that the predetermined leg of the four-legged animal has landed To generate using
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again The measurement program characterized by calculating the average value of the acceleration in the movement of the four-legged animal.

(Supplementary Note 5)
A measuring device attached to a tetrapod,
An acceleration sensor that outputs acceleration data in a three-dimensional space;
A gait identification unit that identifies the gait of the four-legged animal using the acceleration data acquired from the acceleration sensor;
A gait determination unit that determines whether the movement of the right front leg and the left rear leg and the left front leg and the right rear leg of the four-legged animal is paired using the specified gait of the four-legged animal; ,
A kinetic energy value calculating unit that calculates a value indicating kinetic energy for each cycle for three consecutive cycles in the time change of the acquired acceleration data using the acquired acceleration data;
A kinetic energy ratio calculation unit configured to calculate a ratio of kinetic energy between adjacent periods in the three consecutive periods and a ratio of kinetic energy between periods separated by one period using the value indicating the calculated kinetic energy;
Measuring device characterized by having.

(Supplementary Note 6)
A walk determination unit that determines that the walk of the four-legged animal is abnormal when the ratio of kinetic energy between adjacent cycles and the ratio of kinetic energy between cycles separated by one cycle are smaller than a predetermined threshold value; The measuring device according to claim 5, characterized by having.

(Appendix 7)
An index value calculation unit that calculates an index value using an accumulated sum of ratios of kinetic energy between adjacent cycles in the past and an accumulated sum of ratios of kinetic energy between cycles separated by one cycle;
When the cumulative sum of the ratio of kinetic energy between the current adjacent periods and the cumulative sum of the ratio of kinetic energy between the periods separated by one period deviate from the index value by a predetermined amount, the four-legged animal The measurement apparatus according to claim 5, further comprising: a walk determination unit that determines that the walk is abnormal.

(Supplementary Note 8)
A measuring device attached to a tetrapod,
GPS sensor,
An acceleration sensor that outputs acceleration data in a three-dimensional space;
A gait identification unit that identifies the gait of the four-legged animal using the acquired acceleration data;
A gait determination unit that determines whether or not a predetermined leg of the four-legged animal has landed using the specified gait of the four-legged animal;
A position that generates position information indicating the position of the tetrapod when it is determined that the predetermined leg of the tetrapod has landed, using position information and time information indicated by GPS data acquired from the GPS sensor An information generation unit,
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again And an acceleration calculation unit for calculating an average value of accelerations in the movement of the four-legged animal.

(Appendix 9)
Acquire acceleration data in a three-dimensional space from an acceleration sensor attached to a quadruped animal,
Identifying the gait of the tetrapod using the acquired acceleration data;
Using the specified gait of the four-legged animal, it is determined whether or not the movement of the right front leg and the left rear leg, and the left fore leg and the right rear leg of the four-legged animal are paired, respectively.
When it is determined that the movement of the right front leg and the left rear leg, and the left front leg and the right rear leg of the four-legged animal are paired, the temporal change of the acquired acceleration data using the acquired acceleration data Calculate the value indicating the kinetic energy of each cycle for three consecutive cycles in
A measuring method characterized by calculating a ratio of kinetic energy between adjacent cycles in the three consecutive cycles and a ratio of kinetic energy between cycles separated by one cycle using the value indicating the calculated kinetic energy. .

(Supplementary Note 10)
If the ratio of kinetic energy between adjacent cycles and the ratio of kinetic energy between cycles separated by one cycle are smaller than a predetermined threshold, it is determined that the walking of the four-legged animal is abnormal. The measurement method as described in Supplementary Note 9.

(Supplementary Note 11)
An index value is calculated using an accumulated sum of ratios of kinetic energy between adjacent cycles in the past and an accumulated sum of ratios of kinetic energy between cycles separated by one cycle,
When the cumulative sum of the ratio of kinetic energy between the current adjacent periods and the cumulative sum of the ratio of kinetic energy between the periods separated by one period deviate from the index value by a predetermined amount, the four-legged animal 24. The measurement method according to appendix 9, wherein the walking is determined to be abnormal.

(Supplementary Note 12)
Acquire acceleration data in a three-dimensional space from an acceleration sensor attached to a quadruped animal,
Identifying the gait of the tetrapod using the acquired acceleration data;
It is determined whether or not a predetermined leg of the tetrapod animal has landed using the specified gait of the tetrapod animal,
Position information and time information indicated by GPS data acquired from a GPS sensor attached to the four-legged animal, position information indicating the position of the four-legged animal when it is determined that the predetermined leg of the four-legged animal has landed And using
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again Calculating an average value of accelerations in the movement of the four-legged animal.

10 mobile phone terminal 10 ', 10''mobile terminal 104, 304 sub microcomputer 104c, 104h, 104r, 1004a LPF
104 d, 104 o, 104 s HPF
1004b BPF
104e, 104k, 1004d, outer product calculation processing units 104j, 1041, 104m, 104p, 104q, 104t, 104u, 1004c, 1004j, 1004p inner product calculation processing units 10041, 1004n, 1004o multiplication unit 1004m sign determination unit 108 ROM
109 RAM
110 acceleration sensor 111 angular velocity sensor 201 inner product value calculation unit 202 outer product calculation unit 203 absolute value calculation unit 204 operation state identification unit 300, 700 measurement terminal 301 3-axis acceleration sensor 304 b CPU
304c Main memory 304d Flash memory 305 GPS sensor 306 Timer 3001, 7001 Gait identifying unit 3002, 7002 Gait judging unit 3003 Kinetic energy value calculating unit 3004 Kinetic energy ratio calculating unit 7003 Position information generating unit 7004 Acceleration calculating unit

Claims (3)

To the measuring device attached to the tetrapod animal,
Obtaining acceleration data of a three-dimensional space from an acceleration sensor attached to the four-legged animal;
Causing the gait of the tetrapod to be identified using the acquired acceleration data;
It is determined whether or not a predetermined leg of the four-legged animal has landed using the specified gait of the four-legged animal,
Position information and time information indicated by GPS data acquired from a GPS sensor attached to the four-legged animal, position information indicating the position of the four-legged animal when it is determined that the predetermined leg of the four-legged animal has landed To generate using
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again The measurement program characterized by calculating the average value of the acceleration in the movement of the four-legged animal. A measuring device attached to a tetrapod,
GPS sensor,
An acceleration sensor that outputs acceleration data in a three-dimensional space;
A gait identification unit that identifies the gait of the four-legged animal using the acquired acceleration data;
A gait determination unit that determines whether or not a predetermined leg of the four-legged animal has landed using the specified gait of the four-legged animal;
A position that generates position information indicating the position of the tetrapod when it is determined that the predetermined leg of the tetrapod has landed, using position information and time information indicated by GPS data acquired from the GPS sensor An information generation unit,
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again And an acceleration calculation unit for calculating an average value of accelerations in the movement of the four-legged animal. Acquire acceleration data in a three-dimensional space from an acceleration sensor attached to a quadruped animal,
Identifying the gait of the tetrapod using the acquired acceleration data;
It is determined whether or not a predetermined leg of the tetrapod animal has landed using the specified gait of the tetrapod animal,
Position information and time information indicated by GPS data acquired from a GPS sensor attached to the four-legged animal, position information indicating the position of the four-legged animal when it is determined that the predetermined leg of the four-legged animal has landed And using
Until it is determined that the predetermined leg of the four-legged animal has landed using the generated position information and the time information indicated by the GPS data, and then it is determined that the predetermined leg of the four-legged animal has landed again Calculating an average value of accelerations in the movement of the four-legged animal.