JP6529039B2 - Link string mapping device, link string mapping method, and program - Google Patents

Description

  The present invention relates to a link string mapping apparatus and the like used to map a link string of a model to a link string of a robot.

Conventionally, a humanoid robot has been created, and the humanoid robot is also made to imitate human motion. For humanoid robots that can perform movements similar to humans, they can move the captured human movements as they are. On the other hand, if this is not the case, for example, if the humanoid robot has fewer joints than humans, or if the range of motion of the joints is narrow, then it is necessary to convert the captured human motion to suit the robot. The The setting of such transformations has been done manually by humans.
As a related technology, a technology for capturing human motion is known (see, for example, Patent Document 1).

JP 2012-525643 gazette

  As described above, when converting captured human motions to suit the robot, there has been a problem that it is necessary to set the conversion for each type of humanoid robot. In addition, even if such a conversion setting is performed, for example, when a human makes an unexpected movement, the movement of the human can not be properly converted, and the movement of the humanoid robot stops. There was also the problem of

  Generally speaking, when mapping the movement of a model to a robot, there has been a demand that it is not necessary to manually create a conversion equation and the like.

  The present invention has been made in accordance with the above-mentioned circumstances, and an object thereof is to provide a link string mapping device etc. capable of automatically mapping a link string of a model to a link string of any robot.

In order to achieve the above object, the link string mapping apparatus according to the present invention uses a receiving unit that receives model link string information, which is information on the position of each link included in the link string of the model, and model link string information. The purpose according to the distance between the identification part which specifies the coordinate value of a plurality of places determined in advance in the link row of the present invention, the plurality of places identified by the identification part, and the plurality of places in the link row of the robot corresponding to the plurality of places. With respect to the angle of each joint of the link row of the robot according to the robot link row information, the calculation unit that calculates the robot link row information that is information related to the position of each link included in the link row of the robot so that the function becomes smaller. And an output unit for outputting information, wherein the objective function corresponds to each of the weighted distances.
With such a configuration, even if there is a difference, for example, in the number of links, between the link row of the model and the link row of the robot, the link row of the robot is similar to the link row of the model. Information on the angle of each joint can be output. As a result, it is not necessary to manually create a conversion equation or the like for automatically mapping a model link sequence to an arbitrary robot link sequence. In addition, since a conversion formula or the like is not used, robot link string information can be calculated even if the position of the link string of the model becomes unexpected. Therefore, for example, it is possible to avoid a situation in which the movement of the link row of the robot is stopped. In addition, it becomes possible to set which position in the link string of the robot is closer to the position of the link string of the model according to the weight.

Further, in the link string mapping device according to the present invention, the plurality of locations specified by the identifying unit includes the tip of the link string of the model, and the link string of the robot corresponding to the identified location that is the tip of the link string of the model. The location at is the tip of the link row of the robot, and the distance weight of the tip of the link row of each of the model and robot may be larger than the weight of other distances.
Such a configuration makes it possible to approximate positional relationship between the tip of the link row of the model and the tip of the link row of the robot.

Further, in the link string mapping device according to the present invention, the number of the plurality of points may be larger than the number of connection points in the link string of the robot.
Such a configuration makes it possible to properly map the link string of the model to the link string of the robot.

Further, in the link string mapping device according to the present invention, the link string mapping apparatus further includes a storage unit in which joint motion segment information, which is information indicating a motion segment of joints in the link string of the robot, is stored. The robot link string information may be calculated so that the objective function becomes smaller in the region.
With such a configuration, for example, even if the movable range of the joints is different between the link row of the model and the link row of the robot, the robot link row information is made similar to the link row of the model. It can be calculated.

The link string mapping apparatus according to the present invention may further include a normalization unit that normalizes the length of the link string of the model and the length of the link string of the robot, and the distance may be a normalized distance. .
Such a configuration makes it possible to map the link series of the model more appropriately by the link series of the robot.

Further, in the link string mapping device according to the present invention, the distance may be a distance corresponding to a coordinate value in a coordinate system in which the base end of the link string of the model and the base end of the link string of the robot are the origin.
With such a configuration, it is possible to reduce the amount of calculation when calculating robot link string information.

The link string mapping apparatus according to the present invention further includes a noise removing unit that removes high frequency noise components included in the robot link string information, and the output unit is a robot link from which high frequency noise components have been removed by the noise removing unit. Information on the angle of each joint of the link row of the robot according to the row information may be output.
With such a configuration, for example, even when high-frequency noise is included in the model link string information, an output from which the noise is removed can be performed.

The link string mapping apparatus according to the present invention further comprises a prediction unit for performing prediction on model link string information or robot link string information, and the output unit is a link string of a robot predicted according to the prediction result by the prediction unit. The information on the angle of each joint may be output.
With such a configuration, the robot can be operated according to the prediction result, so that, for example, when mapping of link strings is performed in real time, the movement of the robot is not delayed with respect to the movement of the model. can do.

The link string mapping apparatus according to the present invention further includes a prediction unit that predicts model link string information using the model link string information accepted by the accepting unit, and the identifying unit is a model predicted by the predictor. The link string information may be used to specify coordinate values at a plurality of locations.
The link string mapping apparatus according to the present invention further includes a prediction unit that predicts robot link string information using the robot link string information calculated by the calculation unit, and the output unit is a robot predicted by the prediction unit. Information on the angle of each joint of the link row of the robot according to the link row information may be output.
Further, in the link string mapping apparatus according to the present invention, a prediction is performed that predicts information on the angle of each joint using information on the angle of each joint of the link string of the robot according to the robot link string information calculated by the calculation unit. The apparatus may further include a unit, and the output unit may output information on the angle of each joint of the link row of the robot predicted by the prediction unit.

Further, the link string mapping apparatus according to the present invention uses the reception unit for receiving model link string information, which is information on the position of each link included in the link string of the model, and the model link string information, using the model link string information. The objective function according to each distance of the identification part which specifies the coordinate value of the determined plural places, the plural places specified by the identifying part, and the plural places in the link row of the robot corresponding to the plural places becomes small Thus, a calculation unit that calculates robot link string information that is information on the position of each link included in the link string of the robot and information on the angle of each joint of the link string of the robot according to the robot link string information is output An output unit, and a prediction unit that performs prediction on model link string information or robot link string information; Parts are in accordance with the prediction result by the prediction unit, and outputs the information about the angles of the joints of the link row of predicted robot is intended.
With such a configuration, the robot can be operated according to the prediction result, so that, for example, when mapping of link strings is performed in real time, the movement of the robot is not delayed with respect to the movement of the model. can do.

  According to the link string mapping device and the like according to the present invention, it is possible to automatically map the link string of the model to the link string of any robot. As a result, it is not necessary to manually create a conversion equation or the like for mapping. Also, even if the model performs an unexpected operation, robot link string information can be calculated. In addition, when the objective function corresponds to the weight, it is possible to set which position in the link string of the robot is closer to the position of the link string of the model according to the weight. Become. In addition, in the case of performing prediction, for example, even if mapping of a link string is performed in real time, it is possible to prevent the movement of the robot from being delayed with respect to the movement of the model.

Block diagram showing the configuration of a link string mapping apparatus according to an embodiment of the present invention Flow chart showing operation of link string mapping apparatus according to the same embodiment The figure which shows an example of the link sequence of the model in the embodiment The figure which shows an example of the link row of the robot in the embodiment The figure which shows an example of the angle of the joint of the robot in the embodiment A diagram showing an example of a normalized model and a link string of a robot in the same embodiment A diagram showing an example of the operation of a link row of a model and a robot in the same embodiment Block diagram showing another configuration of link string mapping apparatus according to the same embodiment Block diagram showing another configuration of link string mapping apparatus according to the same embodiment Block diagram showing another configuration of link string mapping apparatus according to the same embodiment A schematic view showing an example of the appearance of a computer system according to the embodiment The figure which shows an example of a structure of the computer system in the embodiment.

  Hereinafter, a link string mapping apparatus according to the present invention will be described using an embodiment. In the following embodiments, components and steps denoted by the same reference numerals are the same or correspond to each other, and the description thereof may not be repeated. Since the link string mapping device according to the present embodiment maps the link string of the model to the link string of the robot, the distance between a plurality of points on the link string of the model and a plurality of points on the link string of the robot The angle of each joint of the link row of the robot is obtained so that the corresponding objective function becomes smaller.

  FIG. 1 is a block diagram showing the configuration of a link string mapping apparatus 1 according to the present embodiment. The link string mapping apparatus 1 according to the present embodiment includes a reception unit 11, a storage unit 12, a normalization unit 13, a specification unit 14, a calculation unit 15, a noise removal unit 16, and an output unit 17. .

  The receiving unit 11 receives model link string information which is information on the position of each link included in the link string of the model 5. The model 5 may be, for example, a human or an animal or robot other than human. In addition, since the link string mapping device 1 performs, for example, mapping for imitating the motion of the model 5 to the robot 6, it is preferable that the model 5 has motion. The model 5 is an object to be a mapping source of the mapping of the link string. The model 5 preferably includes one or more link strings to be mapped. For example, when the model 5 is a human, each of the four limbs becomes a link string. Also, it is possible to think of the human head to the waist as one link row. The model link string information may be information of the whole of a plurality of linked links (for example, a link string from the center of both shoulders to the hand, etc.), and a part of the linked plurality of links (for example, It may be information of a link string from the elbow to the hand). Also, the model link string information may be information of a link string having an end such as a hand tip or a toe, that is, an end that is not a joint, or a joint to joint such as a link string from shoulder to wrist. It may be information on the link string up to Further, although in the present embodiment, the case where model link string information is information of a link string not including a branch is mainly described, the model link string information may be information of a link string including a branch. . The model link string information may be any information as long as the information on the position of each link included in the link string of the model 5 can be obtained as a result. The model link string information may be, for example, information indicating the positions of both ends of each link, and may be information indicating the length of each link and the angle between the links in the joint. The information indicating the position of each end of each link may include information indicating the position of each joint, and may further include information indicating the position of the end of the link row. In the present embodiment, a case will be mainly described where model link string information includes information indicating the position of a joint of a link string and information indicating the position of an end. The position is a position in the coordinate system set for the link string of the model 5. The coordinate system may be the coordinate system of the real space in which the model 5 exists, or may be the coordinate system of the virtual space in which the position of the link row of the model 5 is acquired, and other coordinate systems. It is also good.

  The model link string information may be obtained, for example, by the motion capture technology. Motion capture may be performed, for example, by detecting the position of a marker attached to the model 5, or may be performed using a captured image of the model 5. The motion capture performed using a marker may be, for example, an optical motion capture, a mechanical motion capture, or a magnetic motion capture. In addition, when performing motion capture using a captured image, the accuracy of motion capture using the captured image may be improved, for example, by measuring the distance to the model 5 or the like. For example, KINECT (registered trademark) or RealSense may be used as a device for performing motion capture using a captured image.

  When the mapping of the link string is performed in real time, it is preferable that, for example, the receiving unit 11 repeatedly receives the model link string information in time series and receives the same at a fixed or indefinite time interval. That is, it is preferable that the movement of the link string of the model 5 can be known from the model link string information accepted by the accepting unit 11. The model link string information itself may be information along the time series regarding the link string of the model 5. On the other hand, when the mapping of the link string is not performed in real time, the receiving unit 11 may, for example, collectively receive a plurality of model link string information related to the link string of the model 5, or model link string information May be accepted one by one. Further, when the model 5 has two or more link strings, the receiving unit 11 may receive model link string information for each of the link strings of the model 5. In the present embodiment, model 5 is a human being, and model link string information of a link string relating to the human arm is periodically inputted in real time from motion capture device 2 to link string mapping device 1 and accepted by accepting portion 11 Will be mainly described. Further, when the receiving unit 11 receives a plurality of model link string information corresponding to a plurality of link strings, the model link string information includes information identifying the corresponding link string (for example, the link string of the right arm) And the like may be included. The receiving unit 11 may also receive information indicating the orientation of the model 5 (for example, information indicating the orientation of the face of the model 5) as well as the model link string information. By using the information indicating the orientation of the model 5, it is possible to know to which link string of the model 5 the model link string information corresponds.

  The receiving unit 11 may receive, for example, model link string information transmitted via a wired or wireless communication line, and may read it from a predetermined recording medium (for example, an optical disc, a magnetic disc, a semiconductor memory, etc.) The issued model link string information may be accepted. Note that the receiving unit 11 may or may not include a device for receiving (for example, a modem, a network card, etc.). Also, the reception unit 11 may be realized by hardware or may be realized by software such as a driver for driving a predetermined device.

The storage unit 12 stores joint movement range information which is information indicating the movement range of joints in the link row of the robot 6. It is preferable that the joint movement range information includes information on all joints whose movement range is limited. For example, when the angle of the joint Pi between the links 101 and 102 in the link row of the robot 6 is represented by θ ¹ _i , θ ² _i and θ ³ _i as shown in FIG. 5, the joint movement range information is θ ^It may be information indicating a range of one or more angles having a limited range of motion among ¹ _i , θ ² _i and θ ³ _i . For example, the range of the angle θ ¹ _i is
α ¹ _i ≦ θ ¹ _i ≦ β ¹ _i
In this case, the information on the angle θ ¹ _i included in the joint movement range information may be the lower limit value α ¹ _i and the upper limit value β ¹ _i for the angle θ ¹ _i . In addition, about the joint and axis | shaft which do not have a restriction | limiting in a movable area, a movable area does not need to be shown by joint movable area information. Although FIG. 5 shows the case where the joint Pi between the links 101 and 102 has three axes, the number of axes of the joints does not matter. The number of axes of the joints may be, for example, one, two, or three, or four or more. Also, the direction of the axis of the joint is optional. In a joint having two or more axes, the storage unit 12 stores information indicating the order in which the axes are provided, for example, from the link 101 side to the link 102 side. It may be done.

  Although the case where the robot 6 is a humanoid robot is mainly described in the present embodiment, it may not be. The robot 6 may be, for example, an animal robot such as a dog, a cat, a bird, a fish, a snake, or an insect, or may be a robot of another shape. The robot 6 has one or more operable link rows. The operable link row may be, for example, a link row connected by a plurality of joints respectively driven by a plurality of motors. Also, the robot 6 is usually a toy robot, ie, an entertainment robot, but may not be. Further, the number LN1 of links in the link string of the model 5 which is the mapping source and the number LN2 of links in the link string of the robot 6 which is the mapping destination do not matter. For example, the number of links LN1 of the model 5 may be larger or smaller than the number of links LN2 of the robot 6, and both the numbers of links LN1 and LN2 may be equal. In addition, it is assumed that the coordinate system of the link row of the robot 6 is set in the same manner as the link row of the model 5. The coordinate system may be a coordinate system of a real space in which the robot 6 exists, or may be a coordinate system of a virtual space for indicating a positional relationship with respect to a link row of the robot 6, and other coordinate systems. May be

  The storage unit 12 may store information other than the joint movement range information. For example, link length information indicating the length of each link in the link string of the robot 6 may be stored in the storage unit 12. It is known as forward kinematics of a robot that the position of the link row of the robot 6 can be specified by using the link length information and the angle of each joint in the link row of the robot 6. Further, for example, information on the direction of the axis of each joint in the link row of the robot 6 may be stored in the storage unit 12. Further, the storage unit 12 may store information indicating a plurality of predetermined locations in the link string of the model 5 or information indicating a plurality of locations in the link string of the robot 6 corresponding to the plurality of locations. Further, when the link row of the robot 6 includes a link having a shape other than a straight line, information indicating the shape of the link may be stored in the storage unit 12. In addition, a time constant or the like of a low pass filter in the noise removing unit 16 described later may be stored in the storage unit 12.

  There is no limitation on the process of storing information in the storage unit 12. For example, the information may be stored in the storage unit 12 via a recording medium, the information transmitted via a communication line may be stored in the storage unit 12, or The information input via the input device may be stored in the storage unit 12. The storage in the storage unit 12 may be temporary storage in a RAM or the like, or may be long-term storage. The storage unit 12 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, etc.).

  The normalization unit 13 normalizes the length of the link row of the model 5 and the length of the link row of the robot 6. Generally, the length of the link string of the model 5 indicated by the model link string information received from the motion capture device 2 is different from the length of the link string of the robot 6. Therefore, the normalization unit 13 performs normalization, which is processing to match the lengths of the two. In the normalization, the length of the link string of the model 5 and the length of the link string of the robot 6 may be “1” or other values. In the present embodiment, the former case will be mainly described. The length of the link row is the sum of the lengths of links from one end to the other end of the link row. In the case of normalizing the length of the link string of the model 5 and the robot 6 to 1, for example, the normalization unit 13 calculates the length of the link string of the model 5 and calculates each coordinate value in the link string of the model 5 May be normalized by dividing the calculated length of the link string, or the length of each link in the link string of the robot 6 may be calculated, and the length of each link in the link string of the robot 6 may be calculated. Normalization may be performed by dividing by the length of. Note that, for example, when the positions of both ends of each link in the link string of the model 5 are indicated by the model link string information, the normalization unit 13 calculates the length of each link using the information and sums By doing this, the length of the link string of the model 5 can be calculated. Also, for example, when the length and angle of each link in the link string of the model 5 are indicated by the model link string information, the normalization unit 13 sums the lengths of the links, and The length of the link string can be calculated. Further, for example, when the length of each link in the link row of the robot 6 is stored in the storage unit 12, the normalization unit 13 adds the length of each link to calculate the link of the robot 6. The length of the column can be calculated. The normalization unit 13 may perform normalization by changing the unit of at least one of the coordinate system of the link string of the model 5 and the coordinate system of the link string of the robot 6. For example, the normalization unit 13 performs normalization by changing the unit of the coordinate system so that the length of the link string of the model 5 and the length of the link string of the robot 6 become 1 in both coordinate systems. May be

  In the normalization process, the normalization unit 13 may set a coordinate system in which both the base end of the link row of the model 5 and the base end of the link row of the robot 6 are the origin. Then, in the coordinate system, the processing of the identification unit 14 or the calculation unit 15 in the subsequent stage may be performed. Specifically, with respect to each of the coordinate system of the link string of the model 5 and the coordinate system of the link string of the robot 6, the normalization unit 13 coordinates such as parallel movement so that the base end of the link string becomes the origin. In order to convert and make both coordinate systems identical, coordinate conversion such as rotation may be performed on at least one of the coordinate systems. When the directions of both coordinate systems coincide with each other, coordinate conversion such as rotation may not be performed. In the present embodiment, a case where an xyz Cartesian coordinate system is set as a final coordinate system indicating the position of the link row of the model 5 and the robot 6 will be mainly described. The proximal end of the link row is, for example, an end point on the trunk side of the link row corresponding to a human arm or foot, that is, the opposite side (i.e., main body side) of the hand or foot. For example, if there is no body, such as a link row corresponding to a snake, either end, eg, the caudal or cranial end, may be considered as the proximal end. In addition, the end point on the opposite side to the proximal end may be called the distal end. Further, since the length of each link in the link row of the robot 6 does not change, the normalization of the link row of the robot 6 may be performed only once. On the other hand, the length of each link in the link string of the model 5 may change temporally depending on the method of obtaining model link string information. As described above, when the length of each link in the link string of the model 5 can be changed, normalization with respect to the link string of the model 5 is performed each time the receiving unit 11 receives information on the link string of the model 5 Is preferred. On the other hand, when the length of each link in the link string of the model 5 does not change, for example, the normalization unit 13 performs normalization on the link string of the model 5 once, and the coordinate system etc. after normalization is performed. After passing to the motion capture device 2, normalized model link string information may be accepted thereafter. The coordinate values of the model 5 and the robot 6 used by the specifying unit 14 and the calculating unit 15 are coordinate values normalized by the normalizing unit 13. In addition, after normalization is performed and the proximal ends of both link rows are made the origin, it is preferable that the position of the link row of the model 5 and the position of the robot link row are indicated in one coordinate system. is there. In the present embodiment, a case will be mainly described in which the positions of the model 5 and the robot 6 are indicated by the above-described xyz orthogonal coordinate system after normalization is performed and the proximal ends of both link rows are set as the origin.

  The identifying unit 14 uses the model link string information accepted by the accepting unit 11 to identify coordinate values of a plurality of predetermined points in the link string of the model 5. The specification of the coordinate values is specification of the coordinate values after normalization by the normalization unit 13. Further, in the present embodiment, a case where the coordinate value is a coordinate value in the above-mentioned xyz orthogonal coordinate system will be mainly described. A plurality of predetermined locations in the link string of the model 5 are a plurality of locations internally dividing between the end points of the link string of the model 5 at a predetermined ratio. That is, the plurality of locations are locations corresponding to the plurality of points at which both ends of the link string are divided at a predetermined ratio. Therefore, the plurality of locations are determined irrespective of the positions of the joints in the link row. Further, it is preferable that the number of the plurality of places is larger than the number of connection places in the link row of the robot 6. The connection point in the link string is the point at which the link and the link are connected in the link string. Also, since the value obtained by adding 1 to the number of connection points in the link string usually becomes the number of links in the link string, it is preferable that the number of multiple points is equal to or greater than the number of links in the link string of the robot 6 It can be paraphrased that there is. Also, the plurality of positions may or may not be uniformly provided. Further, one or both ends of the link row of the model 5 may or may not be included in the plurality of locations. Further, a plurality of locations internally dividing between the end points of the link row at a predetermined ratio are a plurality of locations internally dividing between the end points at a predetermined ratio when the link row is linearly extended. You may think that it is a place. Each of the plurality of locations may be referred to as a reference point. Further, specifying the coordinate value may be storing the specified coordinate value on the recording medium.

  The calculation unit 15 is configured so that the objective function corresponding to each distance between the plurality of locations whose coordinate values are specified by the identification unit 14 and the plurality of locations in the link row of the robot 6 respectively corresponding to the plurality Robot link string information, which is information on the position of each link included in the link string of 6, is calculated. Here, each of a plurality of points in the link row of the robot 6 corresponding to a plurality of reference points in the link row of the model 5 may also be referred to as a reference point. For example, a plurality of locations in the link array of the robot 6 corresponding to a plurality of locations in the link array of the model 5 identified by the identification unit 14 have, for example, a predetermined ratio between end points of the link array of the robot 6 It may be a plurality of places to internally divide. The predetermined ratio may be the same as or different from the predetermined ratio for a plurality of locations in the link train of model 5. In the latter case, for example, a plurality of reference points in the link string of the model 5 and a plurality of reference points in the link string of the robot 6 so that the motion of the model 5 which is a human becomes an exaggerated motion in the robot 6 May be determined. In any case, when the number of plural places is equal to or more than the number of links in the link row of the robot 6, the ratio is set such that at least one place exists in each link in the link row of the robot 6. May be If the predetermined ratio of the link train of the model 5 and the predetermined ratio of the link train of the robot 6 are the same, the reference of the same ratio in the link train of the model 5 and the link train of the robot 6 The points are the corresponding reference points. On the other hand, when the ratio of the two is different, the plurality of reference points in the link row of the model 5 and the plurality of reference points in the link row of the robot 6 correspond in order from the base end.

  Distances between a plurality of locations specified by the identifying unit 14 and a plurality of locations in the link row of the robot 6 respectively corresponding to the plurality of locations are a specified reference point and a link of the robot 6 corresponding to the reference point It is the distance to the reference point in the column. The objective function according to each distance may be a function whose value increases as each distance increases. Specifically, the objective function may be the sum of squares of each distance. In that case, mapping using the least squares method will be performed. Also, the objective function may be the sum of the absolute value of each distance. In that case, mapping using the least absolute value method will be performed. Needless to say, other objective functions may be used in which the value increases as each distance increases. Each distance is a normalized distance, that is, a distance calculated using normalized coordinate values. Further, each distance is a distance corresponding to a coordinate value in a coordinate system in which a base end of a link row of the model 5 and the robot 6 is an origin, that is, a distance calculated using the coordinate value. Further, calculating the robot link string information so that the objective function becomes smaller means calculating the position of the link string of the robot 6 so as to make the objective function smaller. Usually, since the length of each link in the link row of the robot 6 is constant, the position of the link row of the robot 6 can be specified by the angle of each joint in the link row of the robot 6. Therefore, calculating the position of the link row of the robot 6 may be calculating the angle of each joint in the link row. In that case, the robot link string information may be information indicating the angle of each joint in the link string of the robot 6. If the proximal end of the link row is also a joint, the robot link row information may also include the angle of the joint that is the proximal end. On the other hand, the robot link string information may be information other than the angle of the joint. For example, the robot link string information may be information indicating the positions of both ends of each link in the link string of the robot 6, or may be coordinate values of a plurality of reference points in the link string of the robot 6. When the robot link string information is coordinate values of a plurality of reference points in the link string of the robot 6, it is preferable that the position of the link string can be uniquely identified by the coordinate values of the plurality of reference points. In the present embodiment, a case will be mainly described where the robot link string information is information indicating the angle of each joint in the link string of the robot 6. In that case, in the robot link string information, for example, information for identifying a joint may be associated with information indicating an angle at the joint, and information for identifying an axis of the joint and an angle at the axis The information to show may be matched. Further, when calculating the robot link string information, the position of the base end of the link string of the robot 6 may be fixed or not. In the present embodiment, the former case will be described. In the latter case, information indicating the position of the base end of the link row of the robot 6 may also be included in the robot link row information.

  When calculating robot link sequence information so that an objective function becomes small, the calculation unit 15 can use an arbitrary algorithm for solving an optimization problem. When solving the optimization problem, it may be solved so as to be local optimum or may be solved so as to be global optimum. In the present embodiment, the case of solving the optimization problem so as to be locally optimal will be mainly described. For example, the calculation unit 15 may calculate the robot link string information so that the objective function becomes smaller by the gradient method which is an algorithm using the derivative of the objective function, and the objective function by the algorithm not using the derivative of the objective function The robot link string information may be calculated so that is smaller. Further, the calculation unit 15 calculates the robot link string information so that the objective function becomes smaller in the movable range indicated by the joint movable range information stored in the storage unit 12. That is, the calculation unit 15 may solve the optimization problem with the joint movement range information as a constraint condition. Further, to calculate robot link string information so that the objective function becomes smaller may be to calculate robot link string information so as to minimize the objective function, so that the objective function is close to the minimum The link string information may be calculated. Here, that the objective function is minimized may be that the objective function is strictly the minimum value, or may be that the objective function is sufficiently close to the minimum value. When calculating robot link string information so that the objective function becomes smaller, for example, if it is calculated robot link string information by repeating the iterative process such as the Newton method and ending the processing before completely converging. Although the robot link string information that minimizes the objective function in a strict sense can not be calculated, it can be said that the robot link string information is calculated so that the objective function is sufficiently close to the minimum value. In addition, when the robot link string information is calculated so that the objective function becomes smaller, if the next model link string information is received by the receiving unit 11, the calculating unit 15 determines the robot link string at that time. The information may be passed to the noise removal unit 16 in the subsequent stage, and robot link string information may be calculated according to the newly accepted model link string information. Even in such a case, the calculating unit 15 calculates the robot link string information so that the objective function is close to the minimum. As described above, when the calculation unit 15 ends the process while performing the process sequentially so as to minimize the objective function, the robot link string information at the time of the end becomes smaller in the objective function. It may be considered that it is robot link sequence information calculated as described above.

  The link string of the model 5 and the link string of the robot 6 on which the calculation unit 15 performs the process of reducing the objective function may be determined in advance or may not be predetermined. In the latter case, for example, the link string of the robot 6 corresponding to the link string of the model 5 may be specified according to a predetermined rule. Specifically, in the case where the model 5 is a human and the robot 6 is a humanoid robot, there may be a rule that link strings close to the head correspond to corresponding link strings. There may be a rule in which link strings on the right (or left) relative to head orientation (for example, head orientation) are set as corresponding link strings. This rule makes it possible to associate arm link sequences. In addition, there may be a rule that link rows that are far from the head correspond to corresponding link rows, and link rows that are far from the head and correspond to the right (or left) link direction with respect to the head direction There may be a rule to make it a link string. According to this rule, it is possible to associate foot link strings.

  The noise removing unit 16 removes high frequency noise components included in the robot link string information. For example, when acquiring the model link string information by the motion capture device 2, the model link string information may include high frequency noise components. In that case, as a result, high-frequency noise components are also included in the robot link string information, so it is preferable to remove the noise components by the noise removing unit 16. The noise removing unit 16 may be, for example, a low pass filter. The noise removing unit 16 may function as a low pass filter, for example, by performing moving average processing on the robot link string information calculated by the calculating unit 15. The noise removing unit 16 may read out and use the one stored in the storage unit 12 as the time constant of the low pass filter.

  The output unit 17 outputs information on the angle of each joint of the link row of the robot 6 according to the robot link row information. The robot link string information is robot link string information from which high-frequency noise components have been removed by the noise removing unit 16. When the mapping process is performed in real time, the output unit 17 may repeat the information in time series and may output the information at fixed or indefinite time intervals. In this case, for example, the robot 6 operates in accordance with the movement of the model 5. On the other hand, when the mapping process is not performed in real time, the output unit 17 may sequentially output the information or may collectively output the information. In this case, for example, the robot 6 may reproduce motion similar to the past motion of the model 5. When the robot 6 has two or more link rows, the output unit 17 may output information for each of the link rows of the robot 6. When the robot link string information is information on the angle of each joint of the link string of the robot 6, the output unit 17 may output the robot link string information as it is. When the robot link string information is information indicating the position of the link string of the robot 6, the output unit 17 converts the robot link string information into information on the angle of each joint of the link string of the robot 6 and outputs Do. For example, in FIG. 5, when the positions of the links 101 and 102 are known, a method of specifying the angle with respect to each axis of the joint Pi is already known. It can be converted into information on the angle of each joint of the six link rows. The information on the angle of each joint of the link row of the robot 6 may be information indicating the angle of each joint, and as a result, it may be information on which the angle of each joint can be known.

  Here, this output may be any output as long as the information to be output can be received as a result by the robot 6, for example, even when it is transmitted to the robot 6 via a wired or wireless communication line. It may be stored on a recording medium, or may be delivered to another component. Even when the output unit 17 accumulates information on the recording medium or delivers information to other components, the information finally reaches the robot 6. It is needless to say that is preferable. Moreover, it goes without saying that processing such as format conversion may be appropriately performed until the information to be output is received by the robot 6. Note that the output unit 17 may or may not include a device (eg, a transmitting device) that performs output. Also, the output unit 17 may be realized by hardware, or may be realized by software such as a driver for driving those devices.

Next, the operation of the link string mapping apparatus 1 will be described using the flowchart of FIG.
(Step S101) The normalization unit 13 normalizes the link string of the robot 6.

  (Step S102) The receiving unit 11 determines whether model link string information has been received. If the model link string information is received, the process proceeds to step S103. If not, the process of step S102 is repeated until the model link string information is received.

  (Step S103) The normalization unit 13 normalizes the link string of the model 5 corresponding to the model link string information received in step S102.

  (Step S104) The specifying unit 14 specifies coordinate values at a plurality of predetermined points in the link string of the model 5 normalized in step S103.

  (Step S105) The calculation unit 15 reduces the objective function according to the distances between the plurality of locations whose coordinate values are specified in step S104 and the plurality of locations in the link string of the robot 6 respectively corresponding to the plurality of locations. Then, robot link string information is calculated.

  (Step S106) The noise removal unit 16 removes high frequency noise components included in the robot link string information calculated in step S105.

  (Step S107) The output unit 17 outputs information on the angle of each joint of the link array of the robot 6 according to the robot link array information from which the high frequency noise component has been removed. Then, the process returns to step S102.

  In addition, when performing the process of real time in the flowchart of FIG. 2, when calculating robot link string information in step S105, the process which receives model link string information is performed in parallel and model link string information When the process of step S105 is received, the output corresponding to the robot link string information at that time (steps S106 and S107) is performed, and the process corresponding to the newly accepted model link string information is executed. May be When the model 5 or the robot 6 has a plurality of link strings in a mapping relationship, the link string mapping device 1 repeatedly executes the processing of steps S102 to S107 for each link string. You may run well in parallel. Specifically, when mapping the link rows of both hands and both feet of the model 5 to both hands and feet of the robot 6, the processing of steps S102 to S107 is repeated for four link rows or processed in parallel. It may be done. Further, in the flowchart of FIG. 2, the processing is ended by an interruption of power off or processing end.

  Next, the operation of the link string mapping apparatus 1 according to the present embodiment will be described using a specific example. In this specific example, the case where the link row of the model 5 is a human arm and the link row of the robot 6 is a humanoid robot arm will be described. That is, the case where a human arm is mapped to a humanoid robot arm will be described. In this specific example, the number of links in the link string of the model 5 is n, and the number of links in the link string of the robot 6 is N. n and N each represent an integer of 1 or more.

FIG. 3 is a diagram showing an example of the link string 5a of the model 5 which is a human. Link row 5a to be mapped is composed of links 51 to 53, and the end point p0 which is a joint is the center of both shoulders, the joint p1 is a shoulder, the joint p2 is an elbow, and the end point p3 is a hand is there. In FIG. 3, wrist joints are omitted to simplify the description. 3, in the xyz orthogonal coordinate system respectively the coordinates of _{p0~p3 (x 0, y 0,} z 0), (x 1, y 1, z 1), (x 2, y 2, z 2), ( It is assumed that x ₃ , y ₃ and z ₃ ). The lengths of the links 51 to 53 are d ₁ , d ₂ and d ₃ respectively. In FIG. 3, the number of links n = 3.

FIG. 4 is a view showing an example of the link row 6a of the robot 6 which is a humanoid robot. The link row 6a at the mapping destination is composed of links 61 and 62. The end point P0 which is a joint is a shoulder, the joint P1 is an elbow, and the end point P2 is a hand. In FIG. 4, let coordinates of P0 to P1 be (X ₀ , Y ₀ , Z ₀ ), (X ₁ , Y ₁ , Z ₁ ), and (X ₂ , Y ₂ , Z ₂ ) in the XYZ orthogonal coordinate system, respectively. . The lengths of the links 61 and 62 are D ₁ and D ₂ respectively. It is assumed that D ₁ and D ₂ are stored in the storage unit 12. Further, the angles of the joints at the joints Pi are assumed to be θ ¹ _i , θ ² _i ,..., Θ ^ki _i . Here, ki is an integer of 1 or more indicating the number of axes in the joint Pi. i = 0,1. In FIG. 4, the number of links N = 2.

  Both the xyz orthogonal coordinate system and the XYZ orthogonal coordinate system are local coordinate systems set for the model 5 and the robot 6, and the xyz axes correspond to the XYZ axes, respectively. For example, the z-axis and the z-axis may be set in the body axis direction, the x-axis and the x-axis may be set in the front-rear direction, and the y-axis and the y-axis may be set in the left-right direction. The body axis direction is the vertical direction in the standing position of a human or a humanoid robot.

The normalization unit 13 first normalizes the link string of the robot 6 (step S101). Specifically, the normalization unit 13 uses the length D ₁ or the like of the links stored in the storage unit 12, as shown in the following equation to calculate the length D of the link row.

Next, the normalization unit 13 translates the XYZ orthogonal coordinate system so that P0 becomes the origin. Specifically, (X _i , Y _i , Z _i ) may be set to (X _i -X ₀ , Y _i -Y ₀ , Z _i -Z ₀ ). However, i is 0 to N. Further, in the XYZ orthogonal coordinate system after the parallel movement, the length of the link row of the robot 6 is made to be 1 by multiplying each coordinate value by 1 / D. Specifically, (X _i , Y _i , Z _i ) may be set as (X _i / D, Y _i / D, Z _i / D). The order of the parallel movement process and the process of multiplying 1 / D does not matter. Further, the normalization unit 13, for the length D ₁ or the like of the links stored in the storage unit 12, is normalized by multiplying 1 / D. In the following description, a _{(X i, Y i, Z} i), D i is the coordinate of after performing the translation and normalization, the length of the link.

Thereafter, it is assumed that the receiving unit 11 receives model link string information (step S102). The model link string information is information indicating the positions of both ends of each link in the link string of the model 5, that is, (x ₀ , y ₀ , z ₀ ), (x ₁ , y ₁ , z ₁ ),. It is assumed that x _n , y _n , z _n ). Then, the normalization unit 13 normalizes the link string of the model 5 (step S103). Specifically, the normalization unit 13 calculates the lengths d ₁ to d _n of each link according to the following equation using model link string information.

After that, the normalization unit 13 calculates the length d of the link string using the calculated link length d _{1 and the} like as in the following equation.

Next, the normalization unit 13 translates the xyz Cartesian coordinate system so that p0 becomes the origin. Specifically, (x _i , y _i , z _i ) may be set to (x _i −x ₀ , y _i −y ₀ , z _i −z ₀ ). However, i is 0 to n. Further, in the xyz orthogonal coordinate system after the parallel movement, the length of the link row of the model 5 is made to be 1 by multiplying each coordinate value by 1 / d. Specifically, (x _i , y _i , z _i ) may be (x _i / d, y _i / d, z _i / d). The order of the parallel movement process and the process of multiplying 1 / d does not matter. Also, the normalization unit 13 normalizes the calculated link length d _{1 and the} like by multiplying it by 1 / d. In the following description, it is assumed that (x _i , y _i , z _i ) and d _i are coordinates after translation and normalization and link lengths. In this specific example, since the orientation of the model 5 in the xyz orthogonal coordinate system matches the orientation of the robot 6 in the XYZ orthogonal coordinate system, after performing parallel movement and normalization for each coordinate system, Coordinate values of the model 5 and the robot 6 are shown in the xyz orthogonal coordinate system. That is, parallel movement and the normalized _{(X i, Y i, Z} i) or the like is assumed to be coordinate values in the xyz orthogonal coordinate system.

In this specific example, it is assumed that an objective function according to the distance of the location where the link string of the model 5 and the robot 6 is divided into m is used. That is, it is assumed that a plurality of predetermined locations in the link string of the model 5 are m−1. Further, in the link row of the model 5, the m-1 pieces of position is when the length of the link row and 1, g ₁ is the length from the base end side, g _2, ..., g _m-1 It is assumed that the position is For example, when the link string of the model 5 is equally divided into m, g _i = i / m. Note that g ₀ = 0 and g _m = 1. Further, the coordinates of the reference point corresponding to g _i are (a _i , b _i , c _i ). Also, m is an integer of 2 or more.

In addition, m-1 places in the link line of the robot 6 corresponding to m-1 places in the link line of the model 5 also have a length from the base end side when the length of the link line is 1. Is a position where G ₁ , G ₂ ,..., G _m−1 . For example, when the link string of the robot 6 is equally divided into m, G _i = i / m. Note that G ₀ = 0 and G _m = 1. Further, the coordinates of the reference point corresponding to G _i, and _{(A i, B i, C} i). Also, as noted above, g _i = G _i may or may not be. In addition, g _{1 to} g _m-1 and G _{1 to} G _m-1 may be stored in the storage unit 12.

となるｇ_iが存在する場合に、特定部１４は、そのｉについて、（ｘ_j-1，ｙ_j-1，ｚ_j-1）と、（ｘ_j，ｙ_j，ｚ_j）とを両端とするリンクの長さを１としたときの（ｘ_j-1，ｙ_j-1，ｚ_j-1）からの長さが、

である点を、ｉ番目の基準点（ａ_i，ｂ_i，ｃ_i）とする。そのように、上記式を満たすｇ_iに応じた箇所（ａ_i，ｂ_i，ｃ_i）を特定する処理を、ｊ＝１〜ｎについて行うことによって、モデル５のリンク列におけるあらかじめ決められた複数箇所のそれぞれの座標値を特定することができる。その特定した座標値（ａ_i，ｂ_i，ｃ_i）は、図示しない記録媒体に蓄積されてもよい。なお、ｄ₀＝０としている。 The specifying unit 14 reads g ₁ to g _m-1 from the storage unit 12 and uses them to specify coordinate values at a plurality of locations in the link string of the model 5 (step S104). In particular,

When there exists g _{i which} is, the identifying unit 14 sets both ends of (x _{j -1} , y _{j -1} , z _{j -1} ) and (x _j , y _j , z _j ) for _i. The length from (x _{j -1} , y _{j -1} , z _{j -1} ) when the link length is 1 is

Let the point that is the i-th reference point (a _i , b _i , c _i ). As described above, the process of specifying the portion (a _i , b _i , c _i ) according to g _i satisfying the above equation is performed for j = 1 to n, thereby determining in advance the link string of the model 5 Each coordinate value of multiple places can be specified. The specified coordinate values (a _i , b _i , c _i ) may be stored in a recording medium (not shown). Note that d ₀ = 0.

となるＧ_iが存在する場合に、そのｉについて、（Ｘ_j-1，Ｙ_j-1，Ｚ_j-1）と、（Ｘ_j，Ｙ_j，Ｚ_j）とを両端とするリンクの長さを１としたときの（Ｘ_j-1，Ｙ_j-1，Ｚ_j-1）からの長さが、

である点が、ｉ番目の基準点（Ａ_i，Ｂ_i，Ｃ_i）となる。そのように、上記式を満たすＧ_iに応じた箇所（Ａ_i，Ｂ_i，Ｃ_i）を特定する処理を、ｊ＝１〜Ｎについて行うことによって、ロボット６のリンク列におけるあらかじめ決められた複数箇所のそれぞれの座標値を特定することができる。その特定した座標値（Ａ_i，Ｂ_i，Ｃ_i）は、図示しない記録媒体に蓄積されてもよい。なお、Ｄ₀＝０としている。 Here, a method of specifying the reference points (A _i , B _i , C _i ) in the link row of the robot 6 corresponding to the reference points (a _i , b _i , c _i ) will be described. As with the model 5 link string,

If there exists G _i , then for that i, the length of the link with (X _{j -1} , Y _{j -1} , Z _{j -1} ) and (X _j , Y _j , Z _j ) at both ends The length from (X _{j -1} , Y _{j -1} , Z _{j -1} ) when

Is the i-th reference point (A _i , B _i , C _i ). As such, portions corresponding to the G _i that satisfies the above formula _{(A i, B i, C} i) a process of identifying, by performing the j = 1 to N, are predetermined in the link string of the robot 6 Each coordinate value of multiple places can be specified. The identified coordinate values _{(A i, B i, C} i) can be stored in a recording medium (not shown). Note that D ₀ = 0.

In the link array of the robot 6, the position of the end point on the distal end side of the link can be represented using the position of the end point on the proximal end side of the link and the angle of the joint at the end point on the proximal end side. . For example, it becomes like following Formula. It is to be i = 1 to N, the function F _i, the angle theta ¹ _i of the joint Pi, theta ² _i, ..., a theta ^ki _i as an argument, the function indicating the relative position of P (i + 1) with respect to Pi It is.
(X _i , Y _i , Z _i ) = (X _i-1 , Y _i-1 , Z _i-1 ) + F _i-1 (θ ¹ _i-1 , θ ² _i-1 ,..., Θ ^{k (i -1)} _i-1 )

Therefore, each coordinate value (A _j , B _j , C _j ) can be represented using the angle of each joint Pi, and thus can be represented, for example, by the following equation. Note that j = 1 to m-1. Further, although the arguments of the function G _j may change according to the position of (A _j , B _j , C _j ), the following formula describes the case where the number of arguments is the largest. Since this relationship does not change unless a plurality of points in the link string of the robot 6 is changed, when the identifying unit 14 identifies coordinate values of a plurality of points in the link string of the model 5, a plurality of symbols in the link string of the robot 6 The following equation indicating the location of the point may be calculated, and the calculated result may be stored in the storage unit 12 or a recording medium (not shown).
_{(A j, B j, C} j) = G j (θ 1 0, θ 2 0, ..., θ k0 0, θ 1 1, θ 2 1, ..., θ k1 1, ..., θ 1 N-1, θ ² _N-1 , ..., θ ^{k (N-1)} _N-1 )

Thereafter, the calculation unit 15 calculates robot link string information that minimizes the objective function E (step S105). Specifically, the objective function E of equation, the angle theta ¹ ₀ of each _{^{joint, θ 2 0, ..., θ}} k0 0, θ 1 1, θ 2 1, ..., θ k1 1, ..., θ 1 N _-1 , θ ² _N-1 ,..., Θ ^{k (N-1)} _N-1 is a function having _N-1 as arguments. Therefore, calculation unit 15, the objective function E angle theta ¹ ₀ of each joint as a minimum the following _{^{equation, θ 2 0, ..., θ}} k0 0, θ 1 1, θ 2 1, ..., θ k1 1, ... , Θ ¹ _N−1 , θ ² _N−1 ,..., Θ ^{k (N−1)} _N−1 is calculated, and the robot link string information that is the calculation result is passed to the noise removing unit 16.

The objective function E is, for example, the square of the distance between the reference point ps1 and the reference point PS1, the square of the distance between the reference point ps2 and the reference point PS2, the square of the distance between the reference point ps3 and the reference point PS3 in FIG. The square of the distance between the reference point ps4 and the reference point PS4 is added. If both end points of the link string are also reference points, their distances may be included in the objective function. In that case, in the equation of the objective function E, the sum should be j = 0 to j = m. (A ₀ , b ₀ , c ₀ ) = (x ₀ , y ₀ , z ₀ ), and (a _m , b _m , c _m ) = (x _n , y _n , z _n ), It is assumed that (A ₀ , B ₀ , C ₀ ) = (X ₀ , Y ₀ , Z ₀ ) and (A _m , B _m , C _m ) = (X _N , Y _N , Z _N ). In FIG. 6, since the link trains 5a and 6a after normalization are shown, the lengths of both link trains 5a and 6a are equal, and the end points p0 and P0 on the proximal side coincide with each other. . As described above, when the base ends of the link strings are made to coincide, the distance between the reference points at j = 0 is 0, so in the above objective function, the sum may be j = 1 to j = m. .

In the objective function E, (a _i , b _i , c _i ) are coordinate values specified by the specifying unit 14, and (A _i , B _i , C _i ) is an angle θ ^{1 of} each joint. ^{_{0, θ 2 0, ...,}} θ k0 0, θ 1 1, θ 2 1, ..., θ k1 1, ..., θ 1 N-1, θ 2 N-1, ..., θ k (N-1) N It is represented by _-1 . Therefore, the angle theta ¹ ₀ for the objective function E to a _{^{minimum, θ 2 0, ..., θ}} k0 0, θ 1 1, θ 2 1, ..., θ k1 1, ..., θ 1 N-1, θ 2 N _-1, ..., when determined by theta ^{k (N-1)} gradient _(N-1), for example, by partially differentiating the objective function E by an angle theta ¹ ₀ etc., for the angle theta ¹ ₀ like the objective function E The position at which the local minimum value is obtained may be determined.

  The noise removing unit 16 removes high frequency noise components from the robot link string information received from the calculating unit 15, and passes the robot link string information as a result of the removal to the output unit 17 (step S106). When the noise removal is performed by the moving average process, the noise removing unit 16 may temporarily store a predetermined number of latest robot link string information.

  The output unit 17 outputs the robot link string information received from the noise removing unit 16 to the robot 6 (step S107). As a result, the robot 6 controls the motor or the like of each joint so that each joint of the link array has an angle included in the robot link array information. By doing so, the shape of the link row of the model 5 and the shape of the link row of the robot 6 become similar. Also, by repeating such an operation, the robot 6 can be controlled to imitate the operation of the model 5, and the robot 6 can be operated as a mirror robot.

  An example of the operation of the robot 6 as the mirror robot will be specifically described. As shown in FIG. 7, the model 5 moves the link row 5a corresponding to the arm from the downward position 5a-1 to the upward position 5a-3 via the lateral position 5a-2. Suppose that it moved like arrow AR1. In the series of movements, the arm is untwisted. That is, in the joints p0 and p1, it is assumed that only rotation about an axis perpendicular to the paper surface of FIG. 6 is performed. On the other hand, it is assumed that the link row 6a corresponding to the arm of the robot 6 can be lifted from the downward position 6a-1 without twisting the arm up to the lateral position 6a-2. Then, mapping from the lateral position 6a-2 to the upward position 6a-3 can not be performed without twisting the arm of the robot 6. Therefore, when the link row of the model 5 moves from the side position 5a-2 to the upward position 5a-3, the calculating unit 15 sets the arm 180 in the joint P0 to minimize the objective function E. The robot link string information is calculated to twist. Therefore, when operating the link row 6a of the robot 6 so as to minimize the objective function E, the link row 6a does not twist the arm from the downward position 6a-1 to the side position 6a-2 After moving as indicated by arrow AR2 and then rotating 180 ° in the direction of arrow AR3, it moves from side position 6a-2 to upward position 6a-3 as indicated by arrow AR4. As a result, at the upward positions 5a-3 and 6a-3, the palm of the hand of the model 5 is directed outward, and the palm of the hand of the robot 6 is directed inwardly, but both arms have similar shapes. Can.

  In the above description, when the number of links in the link string of the model 5 of the mapping source is equal to or more than the number of links in the link string of the robot 6, the number and position of plural points in the link string are appropriately set. Thus, the shape of the link row of the robot 6 can be uniquely determined. On the other hand, when the number of links in the link string of the model 5 of the mapping source is less than the number of links in the link string of the robot 6, the shape of the link string of the robot 6 may not be determined uniquely. If the shape of the link string of the robot 6 is not uniquely determined, it is possible to calculate a plurality of different robot link string information that minimize the objective function E. In such a case, the calculation unit 15 may calculate robot link string information at random, for example, or may use robot constraint data to uniquely determine robot link string information. May be calculated. The constraint may be, for example, to minimize the change from the link row of the robot 6 immediately before. The change may be a change in the angle of the joint, or a change in a particular position in the link series. The particular position may be a joint, a tip, or some other point. Further, robot link string information for reducing the objective function E may be calculated using other constraint conditions. Even if the number of links in the link string of the model 5 of the mapping source is equal to or greater than the number of links in the link string of the robot 6, if the shape of the link string of the robot 6 is not uniquely determined, Robot link string information may be calculated.

  As described above, according to the link string mapping apparatus 1 according to the present embodiment, even if a conversion formula or the like for mapping the link string of the model 5 to the link string of the robot 6 is not manually created, The shape of the link string of the model 5 can be mapped to the link string of the robot 6 so that the link string performs an operation similar to that of the model 5 link string. Therefore, there is an advantage that the labor for creating the conversion formula becomes unnecessary. In addition, when such a conversion formula is used, conversion of the robot 6 to a link string may fail due to the model 5 performing an unexpected movement or the like, and the movement of the robot 6 may be stopped halfway. Such a situation can be avoided by performing the mapping as in the present embodiment. Also, if the number of links of the robot 6 is smaller than the number of links of the model 5 or if the range of motion of the joints of the robot 6 is narrower than the range of motion of the joints of the model 5, that is, mapping for complete mirroring is performed. Even if this can not be done, the motion of the robot 6 can be made to resemble the motion of the model 5. Further, by providing the noise removing unit 16, high frequency noise in the robot link string information can be removed, and for example, the influence of high frequency noise mixed in the model link string information can be reduced.

  Also, in the past, human movement was sometimes captured to match the number of links of the humanoid robot and the movement range of the joints. Specifically, when the number of links of the robot arm is one, only the motion related to one link may be captured even in the human arm. In that case, it was necessary to set so as to perform dedicated motion capture for each type of humanoid robot. On the other hand, in the link string mapping device 1 according to the present embodiment, since the mapping from model link string information to robot link string information is automatically performed, the motion capture device 2 is a device that performs general-purpose motion capture. Good. Therefore, there is an advantage that it is not necessary to perform dedicated motion capture.

  In the present embodiment, the link string mapping apparatus 1 includes the noise removing unit 16. However, when the high frequency noise component is not removed, the link string mapping apparatus 1 may use the noise removing unit 16. It does not have to be provided. For example, in the case where it is clear that the model link string information does not include high frequency noise components, the link string mapping device 1 may not include the noise removing unit 16.

  In the present embodiment, the normalizing unit 13 sets the base end of the link row of the model 5 and the base end of the link row of the robot 6 as the origin of the coordinate system. . The robot link string information may be calculated using an objective function in a situation where both base ends are not at the origin, and the coordinate system of each link row is set so that both base ends are at the origin from the beginning. It is also good.

  Moreover, although the case where the length of the link row | line of the model 5 and the length of the link row | line of the robot 6 were normalized was demonstrated in this Embodiment, it may not be so. For example, when both lengths are set to be the same, for example, when the length of the link string in the model link string information and the length of the link string of the robot 6 are determined to be the same. Does not have to process normalization. In that case, the link string mapping device 1 may not have the normalization unit 13.

  Further, in the present embodiment, although the case where the joint movement range information is stored in the storage unit 12 has been described, it may not be. For example, in the case where the range of motion of the joint is not limited, the range of motion information of the joint may not be stored in the storage unit 12. In addition, when the joint movement range information is not stored in the storage unit 12, the calculation unit 15 calculates the robot link string information so that the objective function becomes smaller in the situation without the constraint condition on the joint movement range information. It will be.

  Also, in the present embodiment, as described above, even if the series of processes for calculating robot link string information using model link string information and outputting according to the robot link string information is performed in real time It may or may not be good.

  Moreover, although the case where there is no branch in the link string of the model 5 or the robot 6 has been described in the present embodiment, the link string may include a branch. When the link train of the model 5 or the robot 6 includes a branch, for example, the branch point may be divided into a plurality of link trains, and the above-described processing may be performed for each link train without the branch, or The above-described process may be performed on a link string having

  Moreover, although the case where the mapping for operating the robot 6 as a mirror robot was mainly performed was demonstrated in this Embodiment, it may not be so. For example, the link train of arms of model 5 may be mapped to the link train of legs of robot 6, and the link train of legs of model 5 may be mapped to the link train of arms of robot 6. Also, the link row of the arm of the model 5 that is a human may be mapped to the link row from the head to the tail of the snake-like robot 6.

  In the present embodiment, the case where model link string information is obtained by motion capture has been mainly described, but this may not be the case. The model link string information may be, for example, information created by a manual operation or a computer. In that case, model link string information may be created by 3DCG (three-dimensional computer graphics), for example. As described above, when model link string information created by 3DCG or the like is accepted by the reception unit 11, if model link string information is created for a certain link string, a robot with an arbitrary number of links can be operated It has the merit of becoming

ここで、Ｗ_jは重みであり、例えば、記憶部１２で記憶されていてもよい。その重みＷ_jは、少なくとも１個の値が異なっているものとする。そうでなければ、重みを設定する必要がないからである。また、その重みＷ_jは、ユーザ等が適宜、変更できるようになっていてもよい。そのような場合には、例えば、ユーザが、距離が近くなって欲しい箇所の重みを、他の重みと比較して大きい値に設定することによって、その距離が近くなるようにロボットリンク列情報が算出されるように設定することができる。例えば、特定部１４が特定する複数箇所に、モデル５のリンク列の先端が含まれており、モデル５のリンク列の先端である特定された箇所に対応するロボット６のリンク列における箇所が、そのロボット６のリンク列の先端である場合に、そのモデル５及びロボット６の各リンク列の先端の距離の重みを、他の距離の重みよりも大きく設定することにより、モデル５のリンク列の先端と、ロボット６のリンク列の先端とが近くなるようにロボットリンク列情報が算出されるようにすることができる。具体的には、モデル５及びロボット６の各リンク列の先端の距離の重みを「３」に設定し、他の重みを「１」に設定してもよい。モデル５のリンク列の先端やロボット６のリンク列の先端は、通常、注目される位置であるため、両者が近似した位置関係になることによって、ロボット６の動きがモデル５の動きにより近似したものとなる。また、重みを用いることによってモデル５及びロボット６の各リンク列の所望の位置（例えば、各リンク列の先端の位置）を近づけることができるが、厳密に一致させるのではないため、例えば、関節の可動域等の原因により、所望の位置を一致させると両リンク列の類似性が損なわれる場合には、所望の位置を一致させるよりも、両リンク列の類似性が保たれることが優先されるようになる。その程度については、重みの調整によって設定することができる。 Further, in the present embodiment, the objective function used by the calculation unit 15 in optimization may be one corresponding to each weighted distance. That is, weights are multiplied as coefficients by the distances between the plurality of reference points in the link string of the model 5 and the plurality of reference points in the link string of the robot 6 respectively corresponding to the plurality of reference points included in the objective function It may be Specifically, the objective function E used in the above-mentioned specific example may be changed to the following equation.

Here, W _j is a weight, and may be stored in, for example, the storage unit 12. The weights W _j are different in at least one value. Otherwise, there is no need to set weights. In addition, the weight W _j may be changed appropriately by the user or the like. In such a case, for example, by setting the weight of the portion where the user wants the distance to be close to a large value as compared with other weights, the robot link string information is set so that the distance becomes close. It can be set to be calculated. For example, the tip of the link row of the model 5 is included in a plurality of locations specified by the identifying unit 14, and the location in the link row of the robot 6 corresponding to the identified location that is the tip of the link row of the model 5 is In the case of the tip of the link row of the robot 6, by setting the weight of the distance of the tip of each link row of the model 5 and the robot 6 larger than the weight of the other distances, The robot link string information can be calculated so that the tip and the tip of the link string of the robot 6 are close to each other. Specifically, the weight of the distance of the tip of each link row of the model 5 and the robot 6 may be set to “3”, and the other weight may be set to “1”. Since the tip of the link row of the model 5 and the tip of the link row of the robot 6 are usually at a position of interest, the motion of the robot 6 is approximated by the motion of the model 5 by becoming a positional relationship approximating both It becomes a thing. In addition, although it is possible to bring the desired position of each link row of the model 5 and the robot 6 (for example, the position of the tip of each link row) closer to each other by using the weights, they are not strictly matched. If matching the desired position results in loss of similarity between the two link strings due to reasons such as range of motion, priority is given to maintaining the similarity of both link strings rather than matching the desired positions. Will be The degree can be set by adjusting the weight.

  Further, in the present embodiment, when mapping of a link string is performed in real time, prediction regarding received information and information to be output may be performed, and information according to the prediction may be output. The prediction is to predict information ahead of time (for example, 0.2 seconds, 0.5 seconds, 1 second, etc.). When the robot 6 is operated according to the model link sequence information accepted by the acceptance unit 11, for example, the processing time in the motion capture device 2, the time of transmission from the motion capture device 2 to the link sequence mapping device 1, The operation of the robot 6 is delayed with respect to the operation of the model 5 by the processing time in the link string mapping device 1, the transmission time from the link string mapping device 1 to the robot 6, and the processing time in the robot 6. On the other hand, by predicting model link string information and robot link string information and outputting according to the prediction, the operation of model 5 and the operation of robot 6 can be matched. sell. Therefore, the link string mapping device 1 may include prediction units 21a to 21c as illustrated in FIGS. 8A to 8C. When the prediction units 21 a to 21 c are not particularly distinguished, they are referred to as the prediction unit 21. The prediction unit 21 predicts model link string information or robot link string information. The prediction related to model link string information may be prediction of model link string information. Further, the prediction regarding the robot link string information may be prediction of the robot link string information, or by predicting information about the angle of each joint of the link string of the robot 6 according to the robot link string information It may be. The output unit 17 outputs information on the predicted angle of each joint of the link row of the robot 6 according to the prediction result by the prediction unit 21. For example, when prediction is performed for model link string information, robot link string information and the like calculated according to the prediction result are also predicted, and as a result, the information output by the output unit 17 is also , As expected.

  Here, the method of prediction by the prediction unit 21 will be briefly described. The prediction unit 21 may perform prediction by extrapolation. That is, the prediction unit 21 may predict the information immediately after the predetermined period by extrapolation using the latest prediction target information of the predetermined period. In that case, the extrapolation result is the prediction result. Specifically, the prediction unit 21 uses angles and positions after 200 ms or 400 ms using the latest consecutive predetermined number of angles, positions, etc. (eg, seven angles, positions, etc. every 200 ms latest). May be extrapolated. In addition, the prediction unit 21 may perform prediction by pattern matching. That is, the prediction unit 21 holds the past history of the prediction target information in a recording medium (not shown), and identifies, in the held history, a portion matching the prediction target information of the latest predetermined period, and It may be predicted by reading out the information immediately after the matching point. In that case, the read result is the prediction result. Further, it is preferable that the history held is in time series. Specifically, the prediction unit 21 identifies the location of the history having a high degree of similarity with the latest continuous predetermined number of angles and positions, and reads the angle and position immediately after the identified location, etc. It is also good. Note that having a high degree of similarity may mean that the degree of similarity is higher than a predetermined threshold. In addition, the prediction unit 21 may perform prediction by machine learning. That is, the prediction unit 21 performs machine learning, in which prediction target information of a predetermined period is input and information immediately after the predetermined period is output, using a past history of the prediction target information. Then, using the machine learning result, the prediction unit 21 may perform prediction by outputting information immediately after the predetermined period with respect to the input of prediction target information of the latest predetermined period. In that case, the output information is the prediction result. The prediction target information may be, for example, model link string information, robot link string information, or information on the angle of each joint of the link string of the robot 6 according to the robot link string information It may be Further, immediately after the predetermined period may be, for example, after 0.5 seconds, may be after 1 second, or may be after another time. This time may be set to, for example, the delay time of the motion of the robot 6 with respect to the motion of the model 5 or may be set to another time. When prediction is performed by pattern matching, machine learning, etc., pattern matching, machine learning, etc. is performed for a set of information of a certain link sequence, for example, a set of angles of a plurality of joints of a certain link sequence. May be In addition, the prediction unit 21 may use a plurality of prediction methods in combination. For example, when the prediction by pattern matching can be performed, the prediction unit 21 may perform prediction by pattern matching, and when prediction by pattern matching can not be performed, prediction by extrapolation may be performed. The case where prediction by pattern matching can not be performed may be, for example, a case where there is no history location having high similarity with time changes such as the latest consecutive predetermined number of angles and positions. In addition, the prediction unit 21 may perform prediction by a method other than that described above.

  Next, a case where the link string mapping device 1 includes the prediction unit 21a will be described using FIG. 8A. In this case, the prediction unit 21a predicts model link sequence information using the model link sequence information accepted by the acceptance unit 11. Then, instead of the model link string information received by the receiving unit 11, the model link string information predicted by the prediction unit 21a is used. For example, the specification unit 14 may specify coordinate values at a plurality of locations using the model link string information predicted by the prediction unit 21a. In this case, since the model link string information thus predicted is used, the output target information is also indirectly predicted. Therefore, the output unit 17 outputs information related to the predicted angle of each joint of the link row of the robot 6 according to the prediction result.

  Next, the case where the link string mapping device 1 includes the prediction unit 21b will be described using FIG. 8B. In this case, the prediction unit 21 b predicts robot link sequence information using the robot link sequence information calculated by the calculation unit 15. Then, instead of the robot link string information calculated by the calculating unit 15, the robot link string information predicted by the predicting unit 21b is used. Although FIG. 8B shows the case where the prediction unit 21b predicts robot link string information before noise removal is performed, the robot link after the noise reduction is performed is illustrated. The prediction may be performed on column information. Since the robot link string information predicted in such a manner is used, if the information to be output is the robot link string information itself, the predicted information is to be output, and the information to be output is output If is generated from the robot link string information, indirectly predicted information will be output. Therefore, the output unit 17 outputs information related to the predicted angle of each joint of the link row of the robot 6 according to the prediction result.

  Next, a case where the link string mapping device 1 includes the prediction unit 21c will be described using FIG. 8C. In this case, using the information on the angles of the joints of the link string of the robot 6 according to the robot link string information calculated by the calculating unit 15, the prediction unit 21c predicts the information on the angles of the joints. Do. Then, the output unit 17 outputs information on the angle of each joint of the link row of the robot 6 predicted by the prediction unit 21c. In this case, the output unit 17 outputs information on the angle of each joint of the link row of the predicted robot 6, which is the prediction result itself, according to the prediction result.

  In addition, when prediction is performed on model link string information or robot link string information, the prediction unit 21 may perform prediction on model link string information different from the model link string information received by the receiving unit 11, for example. The prediction may be performed on robot link string information different from the robot link string information calculated by the calculation unit 15. Specifically, when the model link string information accepted by the accepting unit 11 is information indicating the positions of both ends of each link, the prediction unit 21 calculates a link string calculated according to the model link string information. The prediction may be performed on model link string information indicating the angle of the joint of The conversion of the model link string information may be performed by, for example, the prediction unit 21 or may be performed by another component.

  Also, in the above embodiment, each processing or each function may be realized by centralized processing by a single device or a single system, or distributed processing by a plurality of devices or a plurality of systems. It may be realized by

  Further, in the above embodiment, the transfer of information performed between the components is performed by, for example, one of the components if the two components performing the transfer of information are physically different. It may be performed by the output of the information and the reception of the information by the other component, or if the two components that exchange the information are physically the same, one of the components It may be performed by moving from the phase of processing corresponding to to the phase of processing corresponding to the other component.

  Further, in the above embodiment, information related to processing executed by each component, for example, each component receives, acquires, selects, generates, calculates, transmits, receives Even if information such as threshold values, formulas, addresses, etc. used in processing by each component is not specified in the above description, it is held temporarily or for a long time in a recording medium (not shown). May be Further, each component or a storage unit (not shown) may store information in the recording medium (not shown). Each component or a reading unit (not shown) may read information from the recording medium (not shown).

  Further, in the above embodiment, when the information used in each component or the like, for example, information such as a threshold or an address used in processing by each component or various setting values may be changed by the user, Although not explicitly stated in the description, the user may or may not be able to change the information as appropriate. When the user can change such information, the change is realized, for example, by a receiving unit (not shown) that receives a change instruction from the user and a change unit (not shown) that changes the information according to the change instruction. May be The acceptance of the change instruction by the acceptance unit (not shown) may be, for example, acceptance from an input device, reception of information transmitted via a communication line, or acceptance of information read from a predetermined recording medium .

  In the above embodiment, when two or more components included in the link string mapping apparatus 1 have a communication device, an input device, etc., the two or more components have a physically single device. It may also have separate devices.

  Further, in the above embodiment, each component may be configured by dedicated hardware, or a component that can be realized by software may be realized by executing a program. For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. At the time of the execution, the program execution unit may execute the program while accessing the storage unit or the recording medium. The software for realizing the link string mapping apparatus 1 in the above embodiment is the following program. That is, this program uses the computer to receive model link string information, which is information on the position of each link included in the model link string, and the model link string using the model link string information received by the reception unit. The distance between the coordinate value of a plurality of points specified by a specifying unit that specifies the coordinate values of a plurality of points determined in advance in the column, and the coordinate values of a plurality of points in the link row of the robot corresponding to the plurality of places A calculation unit that calculates robot link string information that is information related to the position of each link included in the link string of the robot so that the corresponding objective function becomes smaller, and a robot that corresponds to the robot link string information calculated by the calculation unit Program to function as an output unit that outputs information about the angle of each joint in a link row It is a non.

  In the above program, the function realized by the program does not include the function that can be realized only by hardware. For example, functions that can be realized only by hardware such as an acquisition unit that acquires information, an output unit that outputs information, etc., such as a modem or an interface card are not at least included in the functions that the program realizes.

  In addition, this program may be executed by being downloaded from a server or the like, and the program recorded on a predetermined recording medium (for example, an optical disk such as a CD-ROM, a magnetic disk, a semiconductor memory, etc.) is read. May be performed by Also, this program may be used as a program that constitutes a program product.

  Moreover, the computer that executes this program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

  FIG. 8 is a schematic view showing an example of the external appearance of a computer that executes the program and implements the link string mapping device 1 according to the above embodiment. The above embodiments can be realized by computer hardware and a computer program executed thereon.

  In FIG. 8, a computer system 900 includes a computer 901 including a CD-ROM drive 905, a keyboard 902, a mouse 903 and a monitor 904.

  FIG. 9 is a diagram showing an internal configuration of computer system 900. Referring to FIG. In FIG. 9, a computer 901 is connected to an MPU (Micro Processing Unit) 911, a ROM 912 for storing a program such as a bootup program, and the MPU 911 in addition to a CD-ROM drive 905, and instructions of application programs A RAM 913 for temporarily storing and providing a temporary storage space, a hard disk 914 for storing an application program, a system program, and data, and a bus 915 interconnecting the MPU 911, the ROM 912, and the like are provided. The computer 901 may include a network card (not shown) that provides connection to a LAN, a WAN, or the like.

  The program that causes the computer system 900 to execute the function of the link string mapping device 1 according to the above embodiment may be stored in the CD-ROM 921, inserted into the CD-ROM drive 905, and transferred to the hard disk 914. Alternatively, the program may be transmitted to the computer 901 via a network (not shown) and stored in the hard disk 914. The program is loaded into the RAM 913 upon execution. The program may be loaded directly from the CD-ROM 921 or from the network. Also, the program may be read into the computer system 900 via another recording medium (for example, a DVD) instead of the CD-ROM 921.

  The program may not necessarily include the operating system (OS) that causes the computer 901 to execute the function of the link string mapping apparatus 1 according to the above embodiment, a third party program, or the like. The program may include only portions of instructions that invoke appropriate functions or modules in a controlled manner to achieve a desired result. It is well known how computer system 900 operates, and detailed description is omitted.

  Further, it is needless to say that the present invention is not limited to the above embodiments, and various modifications are possible, which are also included in the scope of the present invention.

  As described above, according to the link string mapping apparatus and the like according to the present invention, an effect of automatically mapping the link string of the model to the link string of any robot is obtained. For example, to operate a humanoid robot or the like as a mirror robot It is useful as an apparatus etc.

DESCRIPTION OF SYMBOLS 1 link row | line | column mapping apparatus 2 motion capture apparatus 11 reception part 12 memory | storage part 13 normalization part 14 identification part 15 calculation part 16 noise removal part 17 output part 21, 21a, 21b, 21c estimation part

Claims (11)

A receiving unit that receives model link string information that is information related to the position of each link included in the model link string;
An identifying unit that identifies coordinate values of a plurality of predetermined points in the link string of the model using model link string information;
Each link included in the link row of the robot such that an objective function corresponding to each distance between the plurality of locations specified by the identifying unit and the plurality of locations in the link sequence of the robot corresponding to the plurality of locations is reduced. A calculation unit that calculates robot link string information that is information related to the position of
An output unit for outputting information on the angle of each joint of the link row of the robot according to the robot link row information;
The link string mapping device, wherein the objective function corresponds to each weighted distance. The tip of the link row of the model is included at a plurality of places specified by the specifying unit,
The location in the link row of the robot corresponding to the identified location that is the tip of the link row of the model is the tip of the link row of the robot,
The link string mapping apparatus according to claim 1, wherein the distance weight of the tip of each link string of the model and the robot is larger than the weight of other distances. The robot further includes a storage unit storing joint movement range information, which is information indicating a movement range of a joint in the link row of the robot.
The link string mapping device according to claim 1 or 2, wherein the calculation unit calculates the robot link string information so that the objective function becomes smaller in a movable range indicated by the joint movement range information. And a normalization unit that normalizes the length of the link string of the model and the length of the link string of the robot,
The link string mapping apparatus according to any one of claims 1 to 3, wherein the distance is a normalized distance. The link according to any one of claims 1 to 4, wherein the distance is a distance according to a coordinate value in a coordinate system in which a base end of the link line of the model and a base end of the link line of the robot are an origin. Column mapping device. A prediction unit that performs prediction on model link string information or robot link string information;
The link string mapping apparatus according to any one of claims 1 to 5, wherein the output unit outputs information on the predicted angle of each joint of the link string of the robot according to the prediction result by the prediction unit. . A receiving unit that receives model link string information that is information related to the position of each link included in the model link string;
An identifying unit that identifies coordinate values of a plurality of predetermined points in the link string of the model using model link string information;
Each link included in the link row of the robot such that an objective function corresponding to each distance between the plurality of locations specified by the identifying unit and the plurality of locations in the link sequence of the robot corresponding to the plurality of locations is reduced. A calculation unit that calculates robot link string information that is information related to the position of
An output unit that outputs information on the angle of each joint of the link row of the robot according to the robot link row information;
A prediction unit that performs prediction on model link string information or robot link string information;
The link string mapping device, wherein the output unit outputs information on the predicted angle of each joint of the link string of the robot according to a prediction result by the prediction unit. An accepting step of accepting model link string information which is information on the position of each link included in the model link string;
An identifying step of identifying coordinate values of a plurality of predetermined points in the link string of the model using the model link string information accepted in the accepting step;
Each link included in the link row of the robot so that an objective function corresponding to each distance between the plurality of locations specified in the identifying step and the plurality of locations in the link row of the robot corresponding to the plurality of locations is reduced. Calculation step of calculating robot link string information which is information related to the position of
Providing an output step of outputting information on the angle of each joint of the link row of the robot according to the robot link row information calculated in the calculation step;
The link string mapping method, wherein the objective function corresponds to each weighted distance. An accepting step of accepting model link string information which is information on the position of each link included in the model link string;
An identifying step of identifying coordinate values of a plurality of predetermined points in the link string of the model using the model link string information accepted in the accepting step;
Each link included in the link row of the robot so that an objective function corresponding to each distance between the plurality of locations specified in the identifying step and the plurality of locations in the link row of the robot corresponding to the plurality of locations is reduced. Calculation step of calculating robot link string information which is information related to the position of
An output step of outputting information on angles of joints of the link row of the robot according to the robot link row information calculated in the calculating step;
Providing a prediction step of predicting model link string information or robot link string information;
The link string mapping method, wherein, in the output step, information related to the predicted angle of each joint of the link string of the robot is output according to the prediction result in the prediction step. Computer,
A reception unit that receives model link string information that is information on the position of each link included in the model link string,
An identification unit that identifies coordinate values of a plurality of predetermined points in the link sequence of the model using the model link sequence information accepted by the acceptance unit;
Each link included in the link row of the robot such that an objective function corresponding to each distance between the plurality of locations specified by the identifying unit and the plurality of locations in the link sequence of the robot corresponding to the plurality of locations is reduced. A calculation unit that calculates robot link string information that is information related to the position of
It functions as an output unit that outputs information related to the angle of each joint of the link row of the robot according to the robot link sequence information calculated by the calculation unit,
The program according to the present invention, wherein the objective function corresponds to each weighted distance. Computer,
A reception unit that receives model link string information that is information on the position of each link included in the model link string,
An identification unit that identifies coordinate values of a plurality of predetermined points in the link sequence of the model using the model link sequence information accepted by the acceptance unit;
Each link included in the link row of the robot such that an objective function corresponding to each distance between the plurality of locations specified by the identifying unit and the plurality of locations in the link sequence of the robot corresponding to the plurality of locations is reduced. A calculation unit that calculates robot link string information that is information related to the position of
An output unit that outputs information on the angle of each joint of the link row of the robot according to the robot link sequence information calculated by the calculation unit;
Function as a prediction unit that performs prediction on model link string information or robot link string information;
The program, wherein the output unit outputs information on the predicted angle of each joint of the link row of the robot according to a prediction result by the prediction unit.

Images