JP6410596B2 - Information processing apparatus and program - Google Patents

Description

  The present invention relates to an information processing apparatus that determines a learning point for learning a pedestrian's stride.

  2. Description of the Related Art Conventionally, navigation devices mounted on a vehicle are known. With the spread of portable terminals such as smartphones, navigation application software can be executed on terminals and function as navigation devices. For this reason, it is possible not only to use a terminal as a car navigation apparatus by mounting the terminal on a car, but also to use it as a pedestrian navigation apparatus when the terminal is moved on foot. In addition, a route guidance system that proposes an optimal combination of various means of transportation such as a train, an airplane, a car, a bus, and a walk from a departure place to a destination has been put into practical use. Therefore, it is required to instruct the next traveling direction at an appropriate timing even when moving on foot.

  For example, in the case of route guidance, the navigation device instructs the next traveling direction slightly before the user changes the traveling direction, but this timing is wrong if the current position of the user is not accurately detected. Even when route guidance is not performed, it is not preferable that the current position of the user and the position of the user on the map differ greatly.

  In a situation where the navigation device can use a Global Navigation Satellite System (GNSS), the current position of the user is detected within a certain error range, so that a relatively accurate current position can be detected. However, in a situation where the navigation device cannot use GNSS, the current position is not necessarily accurately detected because the current position is estimated by autonomous navigation.

  Therefore, a technique for improving the accuracy of the current position at the time of movement by autonomous navigation has been devised (for example, see Patent Document 1). In Patent Document 1, the movement by autonomous navigation is performed when the difference between the position data of the positioning restart point by GPS and the position data measured by autonomous navigation when GPS positioning is resumed is larger than a predetermined value. A positioning device that corrects position data of each point of a moving route by autonomous navigation based on the route length of the route and the route length of the moving route by GPS is disclosed.

JP 2013-76677 A

  However, the positioning device described in Patent Document 1 has a disadvantage that the position data cannot be corrected by autonomous navigation until the radio wave of the GPS satellite can be supplemented.

  Therefore, as a method for correcting the current position detected by autonomous navigation even when GPS satellite radio waves cannot be supplemented, a method using a pedestrian moving between two points whose distances are known is considered.

  FIG. 1 is an example of a diagram illustrating correction of the position of autonomous navigation using the fact that a pedestrian has moved between two points whose distances are known. The terminal 22 includes a sensor unit 12 that transmits an estimated position of a pedestrian and an application 11 that uses the estimated position transmitted by the sensor unit 12.

  As shown in FIG. 1A, the application 11 holds an initial value (an average value or a value set by the user) of the user's step of the terminal 22, and gives this step to the sensor unit 12. . The current location estimating unit 47 of the sensor unit 12 detects walking from vibration or the like, estimates a position separated by a stride from the previous position, and notifies the application 11 of the estimated position. The application 11 can display the position of the user on the map.

  Therefore, if the stride is not accurate, the estimated position estimated by the sensor unit 12 may not be accurate. Therefore, as shown in FIG. 1B, the sensor unit 12 has a stride learning unit 48 as a function of learning the stride. The stride length learning unit 48 acquires an instruction from the application 11 that it is a learning point and a distance between the two points (a distance between the immediately preceding learning point and the current learning point) and learns the user's stride.

  Since the stride length learning unit 48 counts the number of steps from the immediately preceding learning point to the learning point instructed this time, the stride length of the user can be learned by the following equation.

Step = distance between two points / number of steps The current location estimation unit 47 estimates the estimated position using this step, so that a more accurate estimated position can be transmitted to the application 11.

  Here, the application 11 normally corrects the current location obtained by performing map matching or route matching on the estimated position estimated by the sensor unit 12 on the road of the map. Therefore, as a learning point, an intersection where the distance between two points is known can be used, and if a route is already searched (when route matching is performed), a corner on the route can be used. This is because the corner on the route is a place where the user changes the course direction, so that the estimated position is likely to be estimated in the vicinity thereof, and the possibility that the sensor unit 12 performs erroneous learning is low. However, there are the following inconveniences in determining the corner on the route as a learning point.

FIG. 2 is an example of a diagram for explaining inconveniences when turning a corner as a learning point. The marks used in FIG. 2 are as follows.
Solid arrow… Route c1 to destination
Dashed arrow ... Distance learning range c2
Δ: Estimated positions e1 to e15 estimated by the sensor unit 12
○… Current location r1 to r15 after route matching
★ ... Learning points (turning corners on the route) P1, P2
The application 11 corrects (draws) the estimated positions e1 to e15 indicated by Δ on the route by route matching. In general, the correction is made on the route having the shortest distance from the estimated positions e1 to e15. The corrected positions are the current locations r1 to r15.

  As a result of such correction, when the current location (current location r4 in the figure) matches the learning point P1, the application 11 determines that it is a learning point, and determines the distance between the two points (the distance from the departure point to the corner, or the corner). To the next turning corner) is notified to the sensor unit 12. However, there are cases where it is not always appropriate to use a corner as a learning point.

  This is because it is rare that the position on the route closest to the estimated positions e1 to e15 (the estimated position e4 in the figure) coincides with the turning angle, so the timing when the user approaches the turning angle on the route is determined as the learning point. This is because it is difficult to do.

  If the estimated positions e1 to e15 are close to the turning angle on the route to some extent, it may be considered that the estimated position is coincident with the turning angle. However, in this case, since the actual position of the user is considerably away from the learning point, or the number of steps detected by the sensor unit 12 may be significantly different from the actual number of steps, the learning step is not accurate. There is a fear.

  Therefore, when learning the stride using the fact that the distance is moved between two known points, there is a problem that the stride that is learned is not always accurate if the learning point is a turning corner.

  In view of the above problems, an object of the present invention is to provide an information processing apparatus with improved stride learning accuracy.

In view of the above problems, the present invention is an information processing apparatus that estimates the position of a pedestrian using a pedestrian's stride, the position information estimating means for estimating first position information of the pedestrian, and the first If the road where the pedestrian is located changes from the first road to the second road and the angle between the first road and the second road is equal to or greater than the threshold value, it is determined that the pedestrian has changed the course. The position on the second road that is within a predetermined distance after the pedestrian changes the course and does not include the node connecting the first road and the second road is used as a learning point for learning the stride Learning point determination means for determining.

  An information processing apparatus with improved stride length learning accuracy can be provided.

It is an example of the figure explaining correction | amendment of the position of the autonomous navigation using that the pedestrian moved between two points with known distances. It is an example of the figure explaining the inconvenience in the case of using a corner as a learning point. It is a figure which shows an example of the learning point in this embodiment. It is an example of the system block diagram of a route guidance system. It is an example of the hardware block diagram of a navigation server and a terminal. It is an example of the functional block diagram which illustrated each function with which a route guidance system is provided. It is an example of the figure explaining route matching. It is an example of the figure which shows typically determination of the learning point in case the user makes the bent point a learning point. It is an example of the figure which shows the response | compatibility of the step count which a sensor part counts, and a learning point. It is an example of the flowchart figure which shows the operation | movement procedure of a terminal. It is an example of the figure which shows typically determination of the learning point when route matching and map matching are not performed.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

<Schematic features of this embodiment>
FIG. 3 is a diagram illustrating an example of learning points in the present embodiment. One of the features of the application 11 according to the present embodiment is that a user determines a learning point immediately after turning an intersection. In the case of route matching, since the current location obtained by route matching substantially matches the route, it is possible to accurately detect that the user has turned an intersection along the route. Also, if the user actually turns around the intersection, the distance traveled after the turn is short, so the distance between two known points (the distance from the previous learning point to the current learning point) and the user actually moved It can be estimated that the distance is close. Therefore, the sensor unit 12 can learn the user's stride with high accuracy.

<Terms used in this embodiment>
The route matching is to estimate that the position on the route closest to the estimated position estimated by the sensor unit (intersection with the perpendicular line dropped from the estimated position to the route) is the current location.

  Map matching is to estimate that the position on the link closest to the estimated position (intersection with the perpendicular line dropped from the estimated position to the link) is the current location.

  The intersection of this embodiment should just cross | intersect at least two roads, and includes a three-way intersection (T-junction, Y-junction), 4-7 forks, etc. Therefore, the angle at which the user bends depends on the shape of the intersection and is not necessarily 90 degrees.

  The user turning means changing the course along the route or road shape. However, it does not matter whether or not it is detected that the vehicle is along the route or road shape.

  When a user moves a leg forward to move, it is called “walking”.

<Example of system configuration>
FIG. 4 is an example of a system configuration diagram of the route guidance system 100 according to the present embodiment. The route guidance system 100 includes a navigation server 21 and a terminal 22 connected via a network 13. The network 13 includes a communication network constructed mainly by wire such as the Internet, WAN, and LAN, and a communication network constructed mainly wirelessly such as a mobile phone network and a wireless LAN. It may be constructed only by wired or wireless. The terminal 22 connects to the network 13 by accessing the base station 14 and the access point 15 wirelessly.

  The navigation server 21 is an information processing apparatus that provides services and functions related to navigation to the terminal 22. For example, the current location and the destination are acquired from the terminal 22 to search for a route, and route information and a navigation screen described later are transmitted to the terminal 22.

  The terminal 22 is a navigation terminal used by the user. The terminal 22 may be a general-purpose information processing terminal 221 or a navigation dedicated terminal 222. The navigation dedicated terminal 222 is also called a PND (Portable Navigation Device). Note that the terminal 22 of this embodiment may be other than the information processing terminal 221 or the navigation dedicated terminal 222.

  The terminal 22 as the information processing terminal 221 is, for example, a smartphone, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), a notebook PC, or a wearable PC. The information processing terminal 221 is not limited to these and may be any device suitable for route guidance. These devices are normally used as information processing terminals. However, when application software for navigation is executed, route searching, route guidance, and the like are performed as in the case of the navigation dedicated terminal 222.

  In addition, the terminal 22 may be able to switch between an in-vehicle state and a portable state in both cases of the general-purpose information processing terminal 221 and the navigation dedicated terminal 222.

  There are two main modes of operation of the terminal 22. One is a client-type operation mode in which the terminal 22 starts dedicated application software or a Web browser, communicates with the navigation server 21, and receives and displays information on route guidance. The other is an application-type operation terminal that basically completes processing such as drawing a map in the terminal and communicates with the navigation server 21 only when map data acquisition is required. In the present embodiment, a client type is described as an example, but the route guidance of the present embodiment can also be suitably applied to an application type.

  Note that the user may use the route guidance system 100 using the two terminals 22. For example, the terminal 22 such as a notebook PC is accessed to the drive portal site, and a route from the departure place to the destination is searched in advance. The drive portal site is an information service site for drivers. The searched route is registered in the drive portal site, and the route information registered from the terminal 22 such as a smartphone is downloaded at an arbitrary timing.

  FIG. 5 is an example of a hardware configuration diagram of the navigation server 21 and the terminal 22. As shown in FIG. 5A, the navigation server 21 includes a CPU (Central Processing Unit) 211, a ROM (Read Only Memory) 215, a RAM (Random Access Memory) 216, an auxiliary storage device 217, and an input hardware configuration. A device 212, a display device 213, and a communication device 214 are included.

  Further, as shown in FIG. 5B, the terminal 22 includes a CPU 211, a ROM 215, a RAM 216, an auxiliary storage device 217, an input device 212, a display device 213, a communication device 214, a voice input / output device 218, as hardware configurations. And a GPS receiver 219.

  The CPU 211 executes various programs and performs arithmetic processing. The ROM 215 stores programs necessary for startup. The RAM 216 is a work area for temporarily storing processing performed by the CPU 211 and storing data. The auxiliary storage device 217 is a nonvolatile memory that stores various data and programs 2101 and 2102. The input device 212 is, for example, a keyboard or a mouse. The display device 213 is a display or a projector HUD (Head Up Display), and displays, for example, a navigation screen. The communication device 214 is connected to the network 13 and communicates with other devices. The voice input / output device 218 is a device that performs voice input / output. For example, a voice guidance for navigation is output. The GPS receiver 219 is an example of a GNSS (Global Navigation Satellite System) that receives radio waves from a GPS satellite and calculates a current position.

  Note that the input device 212 of the terminal 22 may be realized by a touch panel that can detect a contact position (touch coordinates) on the screen instead of or in addition to the keyboard and the mouse. In addition, the input device 212 may have a function as a speech recognition device that recognizes speech input by the speech input / output device 218.

  The programs 2101 and 2102 stored in the auxiliary storage device 217 of the navigation server 21 or the terminal 22 are distributed in a state of being stored in a storage medium (not shown). Alternatively, it is distributed by downloading from a server (not shown). The program 2102 of the terminal 22 includes the application 11 and further includes software executed by the sensor unit 12. The application 11 may be application software dedicated to route guidance or browser software.

<Functional configuration example of route guidance system>
FIG. 6 is an example of a functional block diagram illustrating each function provided in the route guidance system 100 of the present embodiment. The navigation server 21 includes a server transmission / reception unit 31, a navigation screen creation unit 32, and a route search unit 33. The server transmission / reception unit 31, the navigation screen creation unit 32, and the route search unit 33 are functions or means realized by the CPU 211 shown in FIG. 5 executing the program 2101 and cooperating with the hardware of the navigation server 21. It is. Some or all of these functions may be realized by a hardware circuit such as an IC.

  The navigation server 21 has a map DB 34, a road network DB 35, and a pedestrian network DB 36. These are constructed in, for example, the auxiliary storage device 217 of the navigation server 21, but the navigation server 21 does not have to be directly provided, and may be in a location accessible by the navigation server 21.

  The map DB 34 stores data for drawing the navigation screen. The information displayed on the navigation screen has many display objects such as divisions such as prefectures, green spaces and rivers, roads and railways, symbols and notes, etc., so it can be classified and drawn for each category It is like that. The display object classified into each or the state in which the display object is drawn is called a layer, and the map is drawn by overlapping several layers. The map data of each layer is described in a format suitable for a display target among vector data or raster data. The map data is divided into meshes with known longitudes and latitudes, and a navigation screen is created by combining one or more meshes. In the case of vector beta, the positions of points, polylines, and polygons are determined by latitude and longitude. In the case of raster data, data corresponding to the scale is prepared in association with latitude and longitude.

  The road network DB 35 is data representing the structure of a road through which vehicles can pass, and has a node table and a link table. In the node table, nodes on the road network representation are registered in association with latitude and longitude. A node is called a node. Nodes are, for example, intersections, branch points, merge points, inflection points, and the like. In the link table, roads through which vehicles can pass are registered in association with node numbers of nodes. Roads through which vehicles can pass are ordinary roads, expressways, private roads, private roads, and the like. In the link table, link type, width, link length, and the like are registered. A road between two nodes is called a link, and the link is a line segment connecting the nodes.

  The pedestrian network DB 36 is similar to the road network DB 35 in that it includes a node table and a link table. However, the pedestrian network DB 36 is registered with links of roads (pedestrians, pedestrian crossings, pedestrian bridges, underpasses, passages through which pedestrians and the like) that pedestrians can pass, nodes at the start and end points of the links, and the like.

  In addition, in navigation that proposes an optimal combination of various means of transportation, a route map of a train, a bus operation map, an airplane operation map, and a timetable thereof are used, but they are omitted in the figure.

  The server transmission / reception unit 31 receives various requests related to navigation from the terminal 22. This request includes, for example, a search request to a destination, a navigation screen update request (enlargement / reduction, change of display range, etc.), and the like. These requests are distributed to the navigation screen creation unit 32 and the route search unit 33.

  In response to the search request, the route search unit 33 searches for a route using at least one of the road network DB 35 or the pedestrian network DB 36 and creates route information. The route information includes a link and a node indicating a route from the departure point to the destination, a node for changing the route, node position information, a position for guiding the route change, and the like.

  In the present embodiment, the pedestrian network DB 36 is mainly referred to. For route search, Dijkstra's method is known in which the link length and width are converted into costs, and the route with the smallest total cost from the starting point to the destination is selected. Note that a search method other than the Dijkstra method may be used. In the route search, user settings such as the presence / absence of toll road use and priority for underpass are taken into consideration.

  The route search unit 33 sends the route information from the departure point to the destination obtained by the search to the navigation screen creation unit 32. The navigation screen creation unit 32 creates a navigation screen that includes an area from the departure point to the destination and that highlights the route, the departure point, and the destination. Further, the current position of the user may be displayed. When the user starts moving, the navigation screen creation unit 32 creates a navigation screen with a scale suitable for guidance. Further, when an update request is acquired from the terminal 22, the navigation screen creation unit creates a navigation screen according to the requested scale and display range. The server transmission / reception unit 31 transmits the route information and the navigation screen thus created to the terminal 22.

  The terminal 22 includes a terminal transmission / reception unit 41, an application 11, and a sensor unit 12. The application 11 includes an operation reception unit 42, a route guide unit 43, a navigation screen display unit 44, and a learning point determination unit 45. These are functions or means realized by the CPU 211 shown in FIG. 5 executing the program (application 11) 2102 and cooperating with the hardware of the terminal 22. Some or all of these functions may be realized by a hardware circuit such as an IC.

  The terminal transmission / reception unit 41 transmits a search request and an update request to the navigation server 21 and receives a navigation screen and route information from the navigation server 21.

  The operation reception unit 42 receives operation inputs such as a departure point and a destination, an enlargement / reduction scale, and a display range change from the user. The navigation screen display unit 44 displays a navigation screen on the display device 213. In addition, the estimated position acquired from the sensor unit 12 is corrected on the route (route matching) and displayed on the navigation screen as the current location of the user. In a state where route information has not been searched, map matching for correcting on a link where a user should exist such as a road or a road is performed. Note that the navigation server 21 may perform route matching or map matching.

  The estimated position transmitted by the sensor unit 12 includes a position detected by the GPS receiver 219 and a position estimated by autonomous navigation. In this embodiment, a case will be described in which an estimated position estimated mainly by autonomous navigation is handled.

  The route guidance unit 43 performs route guidance based on the route information acquired from the navigation server 21 and the current location obtained by route matching or map matching. That is, when the user's current location reaches a position to be changed in the route information, the voice input / output device 218 is output to the voice input / output device 218 that indicates a corner. Note that the voice data may be transmitted from the navigation server 21 or may be created by the terminal 22 performing voice synthesis.

  When the learning point determination unit 45 determines that the learning point is a learning point, the learning point determination unit 45 instructs learning point instruction information to the sensor unit 12. The learning point instruction information includes the following information. Details will be described with reference to FIGS.

“Your current location is a learning point, the distance between two points”
Next, the sensor unit 12 will be described. The sensor unit 12 includes an API (Application Interface) 50 serving as an interface with the application 11, a stride learning unit 48, a current location estimation unit 47, and a signal processing unit 46. The stride length learning unit 48, the current location estimation unit 47, and the signal processing unit 46 are functions or means realized by the CPU 211 illustrated in FIG. 5 executing the program 2102 and cooperating with the hardware of the terminal 22. . Some or all of these functions may be realized by a hardware circuit such as an IC.

  Further, the sensor unit 12 has a stride storage unit 49. The stride length storage unit 49 may be built in a flash memory or the like in the sensor unit 12, or may be built in the auxiliary storage device 217 or the ROM 215 of the terminal 22. The stride storage unit 49 stores a stride as an initial value and some strides learned in the past.

  As shown in the drawing, the sensor unit 12 and the application 11 are separated via the API 50, so that the sensor unit 12 can be shared by different applications 11, the sensor unit 12 and the application 11 are developed separately, There is an advantage that even if the part 12 changes, the correction of the application 11 can be minimized. However, the illustrated configuration is an example, and the application 11 and the sensor unit 12 may be integrated.

  The signal processing unit 46 is connected to the acceleration sensor 51, the gyro sensor 52, and the GPS receiver 219. The acceleration sensor 51 is, for example, a sensor that detects acceleration in three directions of the XYZ axes, and the gyro sensor 52 is a sensor that detects angular velocities on planes parallel to the XY plane, the YZ plane, and the ZX plane. Note that the number of axes and the direction of the axes are only examples, and it is only necessary to detect the presence / absence of the user's walking and the traveling direction. The type of sensor that the terminal 22 has is an example.

  The signal processing unit 46 processes signals from the acceleration sensor 51, the gyro sensor 52, and the GPS receiver 219, and sends them to the current location estimation unit 47. When the GPS receiver 219 supplements four or more GPS satellites, the position information of latitude / longitude / elevation detected by the GPS receiver 219 is sent to the current location estimation unit 47. On the other hand, when the GPS receiver 219 cannot capture a sufficient number of GPS satellites, the signals of the acceleration sensor 51 and the gyro sensor 52 are output to the current location estimating unit 47.

  The current location estimating unit 47 monitors signals from the acceleration sensor 51 and the gyro sensor 52 to detect that the user has walked. For example, it is determined based on a signal from the acceleration sensor 51 that the longitudinal acceleration changes by a predetermined amount within a predetermined time or the vertical acceleration becomes a predetermined value or more or less than a predetermined value.

  In addition, the current location estimation unit 47 monitors the signal of the gyro sensor 52 and detects the traveling direction of the user. Since the signal of the gyro sensor 52 is an angular velocity, it is converted into an angle by integrating it. The traveling direction is detected by a change in angle with respect to an arbitrary reference direction (for example, a direction along the route or any one of east, west, south, and north obtained by a geomagnetic sensor).

  The current location estimating unit 47 estimates the current location using the current location immediately before as a base point. The stride recorded in the stride storage unit 49 is read out, and the position advanced by the stride in the traveling direction with respect to the immediately preceding current location is estimated as the current location. The estimated current location is sent to the application 11.

  When acquiring the learning point instruction information from the application 11, the stride learning unit 48 learns the stride using the distance between the two points and the number of steps for moving the distance between the two points. Details will be described with reference to FIGS.

<Route matching>
The learning point determination method of the present embodiment can be applied even if route matching is not performed, but it is easy to determine an accurate learning point by performing route matching.

  FIG. 7A is an example for explaining route matching. A route c1 is obtained by route matching. The estimated positions e1 to e3 estimated by the current location estimating unit 47 are route-matched to the current locations r1 to r3.

  Here, as a route matching procedure, the navigation screen display unit 44 performs map matching prior to route matching. This is because of the logic of the application 11 (map matching is always performed in an area where there is a map). This is because the route matching after correcting the current location by map matching is more accurate.

  For example, the estimated position e4 is once map-matched to the mapping position m4. However, since the mapping position m4 is not on the route, the route is again matched with the current location r4 on the route. If the mapping position m4 remains unchanged, it may be erroneously determined that the user has bent the intersection. However, by performing route matching, it is possible to improve the determination accuracy of whether the user has bent.

  Although the estimated positions e1 to e3 are also map-matched before route matching, route matching can be omitted because the map matching is performed on the route. Therefore, in many cases, even if route matching is performed following map matching, the processing load does not increase greatly.

  In addition, since the estimated position e5 is detected near the intersection, it can be route-matched to either the current location r5a or the current location r5b. However, even if the estimated position e5 is route-matched to the current location r5b before the user actually turns, it is easy to correctly determine whether or not the vehicle is bent based on the location of the current location r6 obtained by subsequent route matching. That is, as shown in FIG. 7B, when the estimated position e6 is route-matched to the current location r6 (or the intersection) before the intersection after the current location r5a or the current location r5b, the user can determine that the vehicle is not bent. . On the other hand, as shown in FIG. 7C, when the estimated position e6 is route-matched to the current location r6 ahead of the intersection, the user can determine that the vehicle is bent.

  Therefore, when route matching is performed, it is possible to use the fact that the user changes the traveling direction along the route, so that it is possible to accurately determine that the user has bent.

  As shown in the figure, when route matching is performed, the intersection is not always closest to the estimated position e5, so the intersection and the current location do not always match. For this reason, it is often not appropriate to make the distance between the intersections a distance between two points, that is, to make the intersection a learning point.

<Determination of learning points When route matching is used>
FIG. 8 is an example of a diagram schematically illustrating determination of a learning point when a point where the user is bent is a learning point, and FIG. 9 is an example of a diagram illustrating correspondence between the number of steps counted by the sensor unit 12 and the learning point. It is. In FIG. 8, the route c1 from the departure point to the destination is searched. This route c1 is composed of links 1 to 4.

  The current location estimation unit 47 processes signals from the acceleration sensor 51 and the gyro sensor 52 to estimate the current location, and sends the estimated locations e1 to e4 to the application 11.

  The navigation screen display unit 44 matches the estimated positions e1 to e4 with the current locations r1 to r4 on the route, and the learning point determination unit 45 of the application 11 acquires the current locations r1 to r4 on the route. The number of steps from the departure point to this point is 4.

  Next, the navigation screen display unit 44 performs route matching with the current location r5 on the route for the estimated position e5 sent by the current location estimation unit 47, and the learning point determination unit 45 of the application 11 acquires the current location r5 on the route. The learning point determination unit 45 detects that the link whose route has been route-matched has changed from the link 1 to the link 2. Whether the link has changed can be determined by whether or not the current location r5 is included between the start node and the end node of link 1.

  When it is detected that the link has changed, the learning point determination unit 45 calculates an angle formed by the link 1 and the link 2, and determines that the user has bent if the formed angle is equal to or greater than a threshold value. Thereby, the learning point can be determined immediately after the user bends. This threshold value is determined depending on how accurately it is possible to detect that the user's traveling direction has changed. Therefore, it is influenced not only by the accuracy of the gyro sensor 52 for detecting the traveling direction but also by how the user holds the terminal 22 and how to walk. However, it is possible to detect a change in the traveling direction of 90 degrees or more at best, and it is possible to detect a change in the traveling direction of about 10 degrees experimentally.

  In the example of FIG. 8, the learning point is immediately after the user bends (the first step after the link is changed), but the learning point may not necessarily be immediately after the user bends. This is because, as described with reference to FIG. 7, incorrect route matching is performed and the user may not actually bend. Assuming such a case, for example, the link on which the user walks changes, and the angle formed by the links may be equal to or greater than a threshold, and the following conditions may be added.

a1. Moved a predetermined distance or more after turning a2. A2. That the current location of a predetermined value or more has been route-matched continuously to the link after being bent. By adding such a condition, it can be detected that the user has bent more accurately. However, the shorter the distance from when the user bends to the determination of the learning point, the easier it is to improve the accuracy of the learned stride.

When the learning point determination unit 45 determines that the user has bent, the learning point determination unit 45 sends learning point instruction information to the stride learning unit 48. The distance between two points included in the learning point instruction information is as follows.
Distance between two points = length of link 1 + distance L ₂₁ (distance from start point of link 2 to current location r5)
The stride learning unit 48 calculates a stride to learn by dividing the distance between two points included in the learning point instruction information by the number of steps. The number of steps from the departure point to the learning point is 5 steps.

Step length = distance between two points / number of steps (5)
Thus, if the learning point is immediately after the user bends, the distance the user has walked is relatively short, so the error is small, and the stride can be calculated with high accuracy.

  In the present embodiment, it has been described that the sensor unit 12 counts the number of steps. However, the application 11 may count the number of steps and notify the sensor unit 12 together with the learning point instruction information.

  When the step length learning unit 48 learns the step length (stored in the step length storage unit 49), the step length learning unit 48 initializes the number of steps to zero and starts counting the number of steps again.

Thereafter, the application 11 starts learning the second stride. That is, when the current location r13 is route-matched, the learning point determination unit 45 determines the current location r13 as a learning point, and sends learning point instruction information to the stride learning unit 48. The distance between the two points at this time is as follows.
Distance between two points = distance L ₂₂ (distance from current location r5 to end point of link 2 + distance of link 3) + distance L ₃₁ (distance from start point of link 4 to current location r13)
The stride learning unit 48 calculates a stride to learn by dividing the distance between two points included in the learning point instruction information by the number of steps. The number of steps from the departure point to the learning point is 8 steps.

Step length = distance between two points / number of steps (8)
Thus, the application 11 repeatedly learns the stride while the user is walking on the route. A fixed number of strides as learning results are retained, and when the determined number is exceeded, the oldest ones are deleted in order. The current location estimating unit 47 estimates the current location using the average or median of the determined number of strides. Therefore, learning accuracy improves as the user walks, and the current location can be estimated with high accuracy. Also, since old learning results are deleted, the current position estimation accuracy can be gradually improved even when the user's stride changes (when it rains, when walking with friends, when the road is congested, etc.).

  In the present embodiment, the stride may be learned from the remaining distance to the destination. For example, when the user reaches the destination, the learning point instruction information is output to the sensor unit 12 with the distance from the current location r13 to the destination as the distance between the two points. In this case, the stride length learning unit 48 can learn the stride regardless of whether or not it is bent. In order to accurately determine that the destination has been reached, it may be determined whether or not the user has stopped near the destination.

<< Terminal Operation Procedure >>
FIG. 10 is an example of a flowchart showing the operation procedure of the terminal. FIG. 10A shows an operation procedure of the application 11, and FIG. 10B shows an operation procedure of the sensor unit 12, respectively.

  The operation procedure of FIG. 10A starts when, for example, the terminal 22 acquires route information from the navigation server 21 and route guidance is started.

  Since the navigation screen display unit 44 performs route matching on the estimated position acquired from the sensor unit 12 to the current location, the learning point determination unit 45 acquires the current location (S10).

  The learning point determination unit 45 identifies the current link where the current location exists (S20). This link may be acquired from the navigation screen display unit 44.

  The learning point determination unit 45 determines whether or not the link where the current location is route-matched has changed (S30). If the link has not changed (No in S30), the process proceeds to step S70.

  When the link has changed (Yes in S30), the learning point determination unit 45 determines whether or not the angle formed by the link before the change and the link after the change is equal to or greater than the threshold (S40). If the formed angle is not equal to or greater than the threshold (No in S40), it can be determined that the user is traveling straight even if the link changes, and the process proceeds to step S70.

  When the formed angle is equal to or greater than the threshold (Yes in S40), it can be determined that the user has bent, so the learning point determination unit 45 calculates the distance between the two points (S50). This distance is the distance from the current location where it was determined to be bent last time to the current location where it was determined to be bent this time.

  Then, the learning point determination unit 45 outputs learning point instruction information to the sensor unit 12 (S60). Thereafter, the process returns to step S10.

  When it is determined No in step S30 or S40, the learning point determination unit 45 determines whether or not the route guidance has ended (S70). If the route guidance has not ended (No in S70), the process returns to step S10.

  When the route guidance ends (Yes in S70), the operation procedure in FIG. 10A ends.

  The operation procedure in FIG. 10B may be started when the application 11 instructs the sensor unit 12 to perform learning, or may be constantly performed.

  For example, when the step length learning unit 48 of the sensor unit 12 acquires an instruction from the application 11 or learning point instruction information, the step number is initialized to zero (S110).

  Next, it is determined whether or not it is determined that the current location estimating unit 47 has walked (S120). When not walking (No in S120), the process proceeds to step S170.

  When it is determined that the user has walked (Yes in S120), the current location estimating unit 47 estimates the estimated position and sends it to the application 11 (S130).

  The stride length learning unit 48 increases the number of steps by one (S140).

  Next, the stride length learning unit 48 determines whether or not learning point instruction information has been acquired for the estimated position transmitted to the application 11 (S150). When learning point instruction information is not acquired (No of S150), since the user is not bent, processing returns to Step S120.

  When the learning point instruction information is acquired (Yes in S150), since the current location is the learning point, the stride learning unit 48 learns the stride by dividing the distance between the two points by the number of steps (S160).

  Thereafter, the process returns to step S110, and the sensor unit 12 continues to learn the stride. And when it determines with No by step S120, the stride learning part 48 determines whether learning was complete | finished (S170). For example, when the stride is learned by an instruction from the application 11, the learning is terminated by an instruction from the application 11.

<Determination of learning points When map matching is performed>
When route matching is not performed and only map matching is performed, a learning point can be determined as in the case where route matching is performed. That is, since the correction destination of the estimated position by the navigation screen display unit 44 is not changed on the route but on the link, the determination of whether or not the vehicle is bent may be the same as in the case of route matching.

  In the case of map matching, since there is a high possibility that the user will be erroneously determined to be bent even if the user is not bent, it is preferable to set the distance of the above condition a1 or the number of walkings of a2 to a larger value.

<Determination of learning points When route matching and map matching are not performed>
There are few cases where even map matching is not performed. In addition, when the estimated position is not stable due to the user's way of holding the terminal 22 or the like, it is not always accurate to determine that the vehicle is bent, and learning of the stride becomes difficult. Therefore, it is preferable that the stride length learning is performed when route matching or map matching is performed. However, when the accuracy of the estimated position is sufficiently high, it is possible to learn the stride even when map matching is not performed.

  FIG. 11 shows an example of a diagram schematically showing determination of learning points when route matching and map matching are not performed. The estimated position output by the sensor unit 12 is almost moved on the link. For this reason, even if it does not correct | amend a user's present position by map matching, it can be considered that it moves along a link.

If route matching and map matching are not performed, the current link cannot be specified. Therefore, whether or not the vehicle is bent is determined by, for example, linear approximation between estimated positions as follows.
(i) The learning point determination unit 45 obtains a straight line passing through two adjacent estimated positions. Thus, a straight line m1 connecting the departure point and the estimated position e1, a straight line m2 connecting the estimated positions e1 and e2, a straight line m3 connecting the estimated positions e2 and e3, a straight line m4 connecting the estimated positions e3 and e4, and the estimated positions e4 and e5 A straight line m5 is obtained.
(ii) The angle formed by the last obtained straight line and the straight line obtained before that is calculated. That is, the angle formed by the straight lines m1 and m2, the angle formed by the straight lines m2 and m3, the straight line m3 and m4, and the angle formed by the straight lines m4 and m5 are obtained.
(iii) If the angle formed is greater than or equal to the threshold, it is determined that the user has bent.
(iv) The learning point instruction information is created by setting the distance between two points as a total value of the straight lines m1 to m5.

  As described above, if the accuracy of the estimated position is sufficient to determine the linearity, it is possible to determine the learning point immediately after the user has bent even when route matching and map matching are not performed.

  Autonomous navigation is performed when the terminal 22 cannot supplement the GPS satellite, but the stride learning may be performed in a state where the GPS satellite is supplemented. In a state where GPS satellites are supplemented, positioning points on the link as shown in FIG. 11 are easily obtained, so even if route matching and map matching are not performed, it is possible to determine immediately after the user is bent as a learning point. Of course, the step length may be learned in a state where the GPS satellite is supplemented and route matching or map matching is performed.

<Preferred modification>
The best mode for carrying out the present invention has been described above with reference to the embodiments. However, the present invention is not limited to these embodiments, and various modifications can be made without departing from the scope of the present invention. And substitutions can be added.

  For example, in this embodiment, the terminal 22 determines whether or not the user has bent, but the navigation server 21 may determine whether or not the user has bent. In this case, the terminal 22 may transmit the estimated position or the current location to the navigation server 21. Further, the sensor unit 12 may determine whether or not the user has bent.

  In the present embodiment, the sensor unit 12 learns the stride, but the application 11 may learn the stride. In this case, the application 11 may count the number of steps between learning points, or may acquire the number of steps from the sensor unit 12.

  Further, the navigation server 21 may perform the learning of the stride. In this case, the navigation server 21 is notified of the fact that the terminal 22 is a learning point, the distance between the two points, and the number of steps. Alternatively, the navigation server 21 may acquire the estimated position or current location, determine that it is a learning point, specify the distance between the two points, and acquire the number of steps from the terminal 22.

  For example, although one navigation server 21 is illustrated in FIG. 4, a plurality of navigation servers 21 may exist. Further, the functions of one navigation server 21 may be distributed and arranged in a plurality of servers.

  Further, the function of the navigation server 21 can be provided on the terminal 22 side. For example, the terminal 22 can have a navigation screen creation unit 32 and a route search unit 33.

DESCRIPTION OF SYMBOLS 11 Application 12 Sensor part 21 Navigation server 22 Terminal 100 Route guidance system

Claims (8)

An information processing apparatus that estimates the position of a pedestrian using a pedestrian's stride,
Position information estimating means for estimating the first position information of the pedestrian;
When the road where the pedestrian is present is changed from the first road to the second road according to the first position information, and the angle between the first road and the second road is equal to or greater than the threshold , the pedestrian has changed the course. And
A position on the second road that is within a predetermined distance after the pedestrian changes course and does not include a node connecting the first road and the second road is determined as a learning point for learning the stride. An information processing apparatus comprising: learning point determination means for performing. Correction means for correcting the first position information to second position information on the road;
The learning point determination unit is configured such that when the road on which the second position information exists changes from the first road to the second road and the angle between the first road and the second road is equal to or greater than a threshold value, the pedestrian The information processing apparatus according to claim 1, wherein the information processing apparatus determines that the course has been changed. When the route information indicating the route to the destination is acquired, the correction unit corrects the first position information to third position information on the route,
The learning point determination means changes the road where the third position information exists from the first road in the route to the second road in the route, and an angle formed by the first road and the second road is a threshold value. The information processing apparatus according to claim 2, wherein the pedestrian determines that the course has been changed in the above case. The information processing apparatus according to claim 1 , wherein the learning point determination unit determines a learning point immediately after the pedestrian changes the course. When the learning point is determined, the learning point determination unit calculates a distance from the immediately preceding learning point to the current learning point and outputs the distance to the stride learning unit, and the distance learning unit The information processing apparatus according to any one of claims 1 to 4 , wherein a step length is learned based on the number of steps between two learning points. The position information estimation means estimates the first position information of a pedestrian using a pedestrian's stride when the first position information cannot be detected by positioning using radio waves from a satellite. Is,
The correction means corrects the first position information to a position on the road, or the correction means corrects the first position information to a position on the route. The information processing apparatus according to claim 3, further comprising display means for displaying the obtained third position information in a map. An information processing apparatus that estimates the position of a pedestrian using a pedestrian's stride,
Position information estimating means for estimating the first position information of the pedestrian;
When the road where the pedestrian is present is changed from the first road to the second road according to the first position information, and the angle between the first road and the second road is equal to or greater than the threshold, the pedestrian has changed the course. And
Learning to learn the pedestrian's stride within a predetermined distance after the pedestrian has changed course and not including the node connecting the first road and the second road on the second road Learning point determination means to determine points,
An information processing apparatus comprising: stride learning means for receiving a command indicating that the learning point is present and learning a stride based on a distance between the two learning points and the number of steps. An information processing device that estimates the position of a pedestrian using the pedestrian's stride
A position information estimation step for estimating the first position information of the pedestrian;
When the road where the pedestrian is present is changed from the first road to the second road according to the first position information, and the angle between the first road and the second road is equal to or greater than the threshold, the pedestrian has changed the course. And
A learning point determination step for determining , as a learning point, a position on the second road that does not include a node that connects the first road and the second road within a predetermined distance after the pedestrian changes the course,
A program that executes

Images