CN-121979396-A - Action simulation method and device, equipment, medium and product

CN121979396ACN 121979396 ACN121979396 ACN 121979396ACN-121979396-A

Abstract

The application provides an action simulation method, an action simulation device, computer equipment, a computer readable storage medium and a computer program product, and belongs to the field of robot control. The motion simulation method comprises the steps of obtaining first motion data of a target object, determining first pose of a plurality of key points in the target object at target moment, optimizing an objective function in a preset joint position range based on the first pose of the plurality of key points and weights corresponding to the plurality of key points, determining a target joint position, wherein the objective function is used for calculating errors between the first pose and second pose of the plurality of key points, the weights are determined based on distribution positions of the corresponding key points, and determining second motion data of the robot based on the target joint position, and the second motion data are used for determining target pose of the plurality of key points in the robot. The technical scheme of the embodiment of the application can realize high-fidelity restoration and accurate expression of action semantics.

Inventors

CHEN CHAO
HUANG WENZHI
HUANG XIAOPENG
DENG GANYU
ZHAO YI
ZHANG JIAXING

Assignees

招商局先进技术开发(深圳)有限公司

Dates

Publication Date: 20260505
Application Date: 20260312

Claims (13)

1. A method of motion simulation, the method comprising: Acquiring first action data of a target object, wherein the first action data are used for determining first pose of a plurality of key points in the target object at a target moment; Optimizing an objective function in a preset joint position range based on the first pose of the plurality of key points and the weights corresponding to the plurality of key points, and determining a target joint position, wherein the preset joint position range comprises a plurality of joint positions used for determining the second pose of the plurality of key points in the robot, the objective function is used for calculating the error between the first pose and the second pose of the plurality of key points, and the weights are determined based on the distribution positions of the corresponding key points; and determining second motion data of the robot based on the target joint positions, wherein the second motion data is used for determining target poses of the plurality of key points in the robot.
2. The method of claim 1, wherein the distribution portion is a key point of the upper limb having a greater weight than other key points.
3. The method of claim 1, wherein the first pose comprises a first position and a first pose, wherein the second pose comprises a second position and a second pose, wherein optimizing the objective function within a range of preset joint positions based on the first pose of the plurality of keypoints and the weights corresponding to the plurality of keypoints, determining the target joint position comprises: Optimizing a first objective function in a preset joint position range based on first poses of the plurality of key points and first weights corresponding to the plurality of key points to obtain an initial joint position, wherein the first objective function is used for calculating errors between the first poses and the second poses of the plurality of key points and errors between the first positions and the second positions of target key points, and the target key points comprise upper limb tail end key points, lower limb tail end key points and root node key points; Optimizing a second objective function in the preset joint position range based on the initial joint position, the first pose of the plurality of key points and the second weight corresponding to the plurality of key points to obtain the target joint position, wherein the second objective function is used for calculating errors between the first pose and the second pose of the plurality of key points, errors between the first position and the second position of the plurality of key points and upper limb pointing loss; wherein the upper limb pointing loss is determined based on a first unit vector in the target object from the root node key point to the upper limb terminal key point and a second unit vector in the robot, and the first weight and the second weight are the same or different.
4. The method of claim 3, wherein the distribution location is a key point of an upper limb, a first weight of the lower limb terminal key point and the root node key point is greater than a first weight of other key points, and a second weight of the key point of the upper limb is greater than a second weight of other key points.
5. The method according to claim 1, wherein the method further comprises: Determining second motion data of the robot at a plurality of continuous moments based on the first motion data of the target object at the plurality of continuous moments, and obtaining an initial motion track of the robot, wherein the initial motion track comprises a plurality of second motion data arranged according to time sequence; processing the initial action track based on a preset strategy to obtain a target action track; Wherein, the difference between the adjacent second motion data in the target motion track is smaller than the difference between the corresponding second motion data in the initial motion track.
6. The method of claim 5, wherein the processing the initial motion trajectory based on the preset strategy to obtain a target motion trajectory comprises: determining abnormal second motion data based on a plurality of second motion data in the initial motion trajectory, wherein the difference between the abnormal second motion data and adjacent second motion data is larger than a first preset threshold; and removing the abnormal second motion data from the initial motion track, and performing interpolation operation to obtain the target motion track.
7. The method of claim 5, wherein the processing the initial motion trajectory based on the preset strategy to obtain a target motion trajectory comprises: determining a joint speed of the robot at each moment based on each of the second motion data in the initial motion trajectory; According to the joint speed at each moment, determining the motion energy of the robot at each moment; Determining a target time interval from the continuous multiple moments based on the motion energy of the robot at each moment, wherein the sum of the motion energy of the robot at each moment in the target time interval is larger than a second preset threshold; and determining the target action track based on the second action data corresponding to the target time interval.
8. The method of claim 7, wherein the determining the target action trajectory based on the second action data corresponding to the target time interval comprises: Acquiring original motion data of the robot, wherein the original motion data corresponds to a preset running state of the robot; Acquiring body constraint data of the robot, wherein the body constraint data comprises speed constraint data; Determining a first current position of a joint based on the original motion data, determining a first target position of the joint based on second motion data corresponding to a first moment in the target time interval, and determining a first interpolation time according to the first current position of the joint, the first target position of the joint and the speed constraint data; Determining a second current position of a joint based on second motion data corresponding to the last moment in the target time interval, determining a second target position of the joint based on the original motion data, and determining a second interpolation time according to the second current position of the joint, the second target position of the joint and the speed constraint data; And splicing the first transition action track, the second action data corresponding to the target time interval and the second transition action track to obtain the target action track.
9. The method of claim 8, wherein the stitching the first transitional motion profile, the second motion data corresponding to the target time interval, and the second transitional motion profile to obtain the target motion profile comprises: Splicing the first transition action track, the second action data corresponding to the target time interval and the second transition action track to obtain a spliced action track; Optimizing a third objective function based on a preset constraint condition and the splicing action track to obtain the objective action track; The third objective function is used for calculating the difference, joint speed and joint acceleration between the joint position and the reference joint position at each moment, the reference joint position is determined based on the spliced action track, and the preset constraint conditions comprise a joint position range, a joint speed range and a joint acceleration range, and joint positions and joint speeds at the track start point and the track end point.
10. An action simulation device, the device comprising: The system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring first action data of a target object, and the first action data are used for determining first pose of a plurality of key points in the target object at a target moment; The optimization module is used for optimizing an objective function in a preset joint position range based on the first pose of the plurality of key points and the weight corresponding to each of the plurality of key points to determine a target joint position, wherein the preset joint position range comprises a plurality of joint positions, the joint positions are used for determining the second pose of the plurality of key points in the robot, the objective function is used for calculating the error between the first pose and the second pose of the plurality of key points, and the weight is determined based on the distribution position of the corresponding key points; The determining module is used for determining second action data of the robot based on the target joint position, wherein the second action data are used for determining target poses of the plurality of key points in the robot.
11. A computer device, comprising: At least one processor, and A memory communicatively coupled to the at least one processor, wherein: the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 9.
12. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein computer instructions which, when executed by a processor, implement the method of any of claims 1 to 9.
13. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 9.

Description

Action simulation method and device, equipment, medium and product Technical Field The present application relates to the field of robot control technology, and in particular, to a motion simulation method, a motion simulation apparatus, a computer device, a computer readable storage medium, and a computer program product. Background With the increasing demand of public places (such as museums and shops) for intelligent navigation and interaction services, service robots are gradually upgraded from simple mobile chassis to intelligent terminals with interaction capability, so that the robots can move and simulate human beings to make corresponding limb actions. However, the current motion simulation technology still has difficulty in realizing high-fidelity restoration of human motions and inaccurate motion semantic expression, so that human-computer interaction experience is poor. It should be noted that the foregoing is not necessarily prior art, and is not intended to limit the scope of the present application. Disclosure of Invention The embodiment of the application provides an action simulation method, an action simulation device, computer equipment, a computer readable storage medium and a computer program product, which enable the process of mapping human action data into robot action data in a cross-form manner to pay more attention to action semantics of specific parts by distributing different weights to key points of different distribution positions, so that high-fidelity reduction and accurate expression of the action semantics can be realized on the basis of considering actual application differences of all parts of a robot, and man-machine interaction experience is improved. In a first aspect, an embodiment of the present application provides an action simulation method, including: Acquiring first action data of a target object, wherein the first action data are used for determining first pose of a plurality of key points in the target object at a target moment; Optimizing an objective function in a preset joint position range based on the first pose of the plurality of key points and the weights corresponding to the plurality of key points, and determining a target joint position, wherein the preset joint position range comprises a plurality of joint positions used for determining the second pose of the plurality of key points in the robot, the objective function is used for calculating the error between the first pose and the second pose of the plurality of key points, and the weights are determined based on the distribution positions of the corresponding key points; and determining second motion data of the robot based on the target joint positions, wherein the second motion data is used for determining target poses of the plurality of key points in the robot. In a second aspect, embodiments of the present application provide an action simulation apparatus, the apparatus comprising: The system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring first action data of a target object, and the first action data are used for determining first pose of a plurality of key points in the target object at a target moment; The optimization module is used for optimizing an objective function in a preset joint position range based on the first pose of the plurality of key points and the weight corresponding to each of the plurality of key points to determine a target joint position, wherein the preset joint position range comprises a plurality of joint positions, the joint positions are used for determining the second pose of the plurality of key points in the robot, the objective function is used for calculating the error between the first pose and the second pose of the plurality of key points, and the weight is determined based on the distribution position of the corresponding key points; The determining module is used for determining second action data of the robot based on the target joint position, wherein the second action data are used for determining target poses of the plurality of key points in the robot. In a third aspect, an embodiment of the present application provides a computer apparatus, including: At least one processor, and A memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as described above. In a fourth aspect, embodiments of the present application provide a computer readable storage medium having stored therein computer instructions which, when executed by a processor, implement a method as described above. In a fifth aspect, embodiments of the present application provide a computer program product comprising a computer program which, when executed by a processor, implements a method as described above. In the technical scheme provided by some embodimen