Program, information processor, and information processing method

1. A non-transitory computer-readable medium having embodied thereon a program, which when executed by a computer causes the computer to execute an information processing method, the method comprising:

implementing a correction function of referencing a first position information obtained by performing a first process on sensor data acquired by two or more motion sensors attached to two or more sites of a body and a second position information estimated by using a learning model on the sensor data;

correcting the first position information for attachment sites to which the motion sensors are attached on a basis of the second position information; and

implementing an interpolation function of estimating, on a basis of the first position information for the attachment sites corrected by the correction function, position information for a non-attachment site of the body to which none of the motion sensors are attached, the position information for the non-attachment site not including the first position information for the attachment sites,

wherein the first position information and the second position information include coordinates of the attachment sites in a coordinate system and the position information for the non-attachment site of the body include coordinates of the non-attachment site in the coordinate system.

2. The non-transitory computer-readable medium according to claim 1 , wherein

the sensor data includes an acceleration,

the first process includes integration of the acceleration, and

the first position information for the attachment sites is obtained by the integration.

3. The non-transitory computer-readable medium according to claim 1 , wherein

the sensor data includes an acceleration, and

the second position information is estimated in regression by using the learning model based on the acceleration.

4. The non-transitory computer-readable medium according to claim 3 , wherein

the sensor data further includes an angular velocity, and

the second position information is estimated in regression by using the learning model further based on the angular velocity.

5. The non-transitory computer-readable medium according to claim 4 , wherein the acceleration, after being subjected to a coordinate system conversion based on the angular velocity, is used for estimation in regression.

6. The non-transitory computer-readable medium according to claim 4 , wherein orientation information for the attachment sites is obtained based on the angular velocity and is used for estimation in regression.

7. The non-transitory computer-readable medium according to claim 4 , wherein the acceleration or the angular velocity acquired by the motion sensors is used for estimation in regression.

8. The non-transitory computer-readable medium according to claim 3 , wherein the estimation in regression is performed on a basis of position information for the attachment sites previously performed by the estimation in regression.

9. The non-transitory computer-readable medium according to claim 3 , wherein an interpolation process estimates in regression, on a basis of the second position information for the attachment sites estimated in regression, position information or orientation information for a non-attachment site to which none of the motion sensors are attached.

10. The non-transitory computer-readable medium according to claim 9 , wherein the correction function further corrects the position information or the orientation information for the non-attachment site included in the second position information on a basis of the first position information.

11. The non-transitory computer-readable medium according to claim 1 , wherein the executed method further comprises:

implementing a reliability identification function of identifying reliability that indicates reliability of the position information for the attachment sites related to the first positional information or the second positional information.

12. The non-transitory computer-readable medium according to claim 11 , wherein the correction function corrects the first position information for the attachment sites on a basis of the reliability identified by the reliability identification function.

13. The non-transitory computer-readable medium according to claim 1 , wherein the interpolation function estimates the position information or orientation information for the non-attachment site located between a plurality of the attachment sites in a predetermined skeleton structure.

14. The non-transitory computer-readable medium according to claim 13 , wherein the interpolation function estimates the position information or the orientation information for the non-attachment site by regression.

15. The non-transitory computer-readable medium according to claim 1 , wherein the correction function is implemented by a Kalman filter.

16. The non-transitory computer-readable medium according to claim 1 , wherein the correction function detects a ground contact state and corrects the first position information for the attachment sites on a basis of a detection result.

17. The non-transitory computer-readable medium according to claim 16 , wherein the correction function estimates a velocity and a position of a forefoot using an acceleration and an angular velocity included in the sensor data, and detects the ground contact state on a basis of an estimation result, the ground contact state being a state in which the forefoot has a predetermined velocity.

18. An information processor comprising

a corrector that references a first position information obtained by performing a first process on sensor data acquired by two or more motion sensors attached to two or more sites of a body and a second position information estimated by using a learning model on the sensor data, corrects the first position information for attachment sites to which the motion sensors are attached on a basis of the second position information, and implements an interpolation function of estimating, on a basis of the first position information for the attachment sites corrected by the correction function, position information for a non-attachment site of the body to which none of the motion sensors are attached, the position information for the non-attachment site not including the first position information for the attachment sites,

wherein the first position information and the second position information include coordinates of the attachment sites in a coordinate system and the position information for the non-attachment site of the body include coordinates of the non-attachment site in the coordinate system.

19. An information processing method comprising:

referencing, by a processor, a first position information obtained by performing a first process on sensor data acquired by two or more motion sensors attached to two or more sites of a body and a second position information estimated by using a learning model on the sensor data;

correcting, by the processor, the first position information for attachment sites to which the motion sensors are attached on a basis of the second position information; and

implementing an interpolation function of estimating, on a basis of the first position information for the attachment sites corrected by the correction function, position information for a non-attachment site of the body to which none of the motion sensors are attached, the position information for the non-attachment site not including the first position information for the attachment sites,

wherein the first position information and the second position information include coordinates of the attachment sites in a coordinate system and the position information for the non-attachment site of the body include coordinates of the non-attachment site in the coordinate system.