CN-121756357-B - Motion object control method and system based on motion capture

CN121756357BCN 121756357 BCN121756357 BCN 121756357BCN-121756357-B

Abstract

The invention provides a motion object control method and a motion object control system based on motion capture, wherein the method comprises the steps of acquiring whole body gesture data of a target operator acquired by a motion capture device, converting the whole body gesture data into a target position sequence corresponding to a target motion object coordinate system, performing redundant motion detection on the target position sequence, performing self-adaptive smoothing on the target position sequence according to the result of the redundant motion detection, thereby obtaining a smooth object control instruction, inputting the smooth object control instruction into a driving controller of a target motion object, and further controlling the target motion object to reduce unnecessary driving motion and reduce energy consumption when executing capture motion.

Inventors

Zhu Zonghuang
YANG YUNJI
LIU JUNHAO
WANG BIN

Assignees

南京因克斯智能科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260302

Claims (6)

1. A motion object control method based on motion capture, comprising: Acquiring whole body posture data of a target operator acquired by a motion capture device; Converting the whole body attitude data into a coordinate system to obtain an object target position sequence corresponding to a target moving object coordinate system; redundant motion detection is carried out on the object target position sequence, and adaptive smoothing is carried out on the object target position sequence according to the result of the redundant motion detection, so that a smooth object control instruction is obtained; Inputting the smoothing object control instruction to a drive controller of the target moving object; the redundant motion detection of the target position sequence of the object comprises the following steps: determining a high-frequency component higher than a preset cut-off frequency in the target position sequence of the object based on frequency domain analysis; judging whether the object position change amplitude corresponding to the high-frequency component is smaller than a preset object inching threshold value or not based on an amplitude threshold value; the adaptive smoothing processing of the target position sequence of the object comprises the following steps: performing self-adaptive Kalman filtering processing on the target position sequence of the object to dynamically adjust the proportion of process noise and measurement noise according to real-time noise estimation; performing velocity smoothing interpolation based on a third-order polynomial on the filtered object position data to limit the object acceleration change rate to be smaller than a preset acceleration threshold value; Before the velocity smoothing interpolation based on the third-order polynomial is performed on the filtered object position data, the method further comprises: Constructing a second-order kinematic prediction model based on the object position, the object speed and the object acceleration so as to predict the future object target position according to a preset delay compensation time, wherein the delay compensation time corresponds to a phase delay introduced by the interpolation window length adopted by the speed smooth interpolation; the construction of a second-order kinematic prediction model based on the object position, the object speed and the object acceleration to predict the future object target position according to a preset delay compensation time comprises the following steps: acquiring a recent historical sampling value of the object position and corresponding object speed and object acceleration; The object speed and the object acceleration are subjected to second-order Taylor expansion on the basis of the object position, and an instantaneous predicted value of the future object position is obtained; and inputting the preset delay compensation time into the second-order Taylor expansion calculation formula as a prediction time step to generate a future object target position corresponding to the preset delay compensation time.
2. The motion capture-based moving object control method of claim 1, wherein the inputting the smoothing object control instruction to the target moving object drive controller comprises: A drive feedforward component based on a predicted object speed is configured in the smoothed object control command to compensate for dynamic response hysteresis of the target moving object due to inertia and friction damping.
3. The motion capture-based moving object control method according to claim 2, wherein configuring a predicted object speed-based drive feedforward component in the smoothing object control instruction comprises: determining a target driving moment corresponding to the speed of the predicted object according to the second-order kinematic prediction model; and linearly superposing the target driving torque and the object position control quantity of the smooth object control command to obtain a combined control command containing the object position control quantity and the driving feedforward component.
4. The motion capture-based moving object control method of claim 3, wherein the determining a target drive torque corresponding to the predicted object speed from the second-order kinematic prediction model comprises: acquiring the instantaneous predicted angular velocity and the instantaneous predicted angular acceleration of the target moving object; Calculating an inertial driving moment based on the equivalent moment of inertia parameter of the target moving object; calculating friction driving moment based on the viscous friction coefficient of the target moving object; And carrying out weighted summation on the inertia driving moment and the friction driving moment to obtain the target driving moment.
5. The motion capture-based moving object control method of claim 3, wherein the linearly superimposing the target drive torque and the object position control amount of the smoothing object control instruction comprises: Converting the object position control quantity into a position ring output torque, and adopting the target driving torque as a speed ring output torque; linearly superposing the position loop output torque and the speed loop output torque in the same object control channel according to a preset weight coefficient to obtain a combined control torque; And the driving controller is used for converting the combined control moment into the combined control command and inputting the combined control command into the target moving object.
6. A motion capture based motion object control system for performing the motion capture based motion object control method of any one of claims 1-5, the system comprising: The acquisition module is used for acquiring the whole body posture data of the target operator acquired by the motion capture device; The processing module is used for converting the whole body posture data into a coordinate system to obtain an object target position sequence corresponding to a target moving object coordinate system; the processing module is used for carrying out redundant motion detection on the object target position sequence, and carrying out self-adaptive smoothing processing on the object target position sequence according to the result of the redundant motion detection to obtain a smooth object control instruction; And the control module is used for inputting the smooth object control instruction to the driving controller of the target moving object.

Description

Motion object control method and system based on motion capture Technical Field The present application relates to a device control technology, and in particular, to a motion object control method and system based on motion capture. Background With the development of man-machine interaction and intelligent manufacturing fields, motion object control methods based on motion capture are widely used in systems such as humanoid robots, collaborative robots, remotely operated mechanical arms and the like. The existing motion capture device (comprising an optical capture system, an inertial measurement unit, a magnetic sensing system and the like) can acquire whole body posture data of a target operator in real time, and map the captured posture information into a joint position instruction under a corresponding robot joint coordinate system through coordinate system conversion. The control method can realize the real-time restoration of the actions of an operator, so that the robot realizes the reproduction of the highly simulated human body actions, and has stronger application potential in the scenes of entertainment, medical rehabilitation, virtual reality, industrial production and the like. However, in the prior art, since the acquisition process of the motion capture device is inevitably affected by the non-purposeful micro-motion of the operator, the resulting sequence of joint target positions often contains redundant motion data that is relatively frequent, relatively small in magnitude, and independent of the actual task target. These redundant data are directly input to the robot joint drive controller without being processed, which results in the robot joint performing a large number of meaningless small amplitude drive actions. This not only increases the inefficient power consumption of the drive system, but may also cause micro-jitter in the joint pose, even affecting the endurance of the robot in long-term operation and the service life of the joint mechanical parts. Disclosure of Invention The application provides a motion object control method and a motion object control system based on motion capture, which are used for controlling a target robot joint to reduce unnecessary joint driving motion and reduce energy consumption when the target robot joint executes capture motion. In a first aspect, the present application provides a motion object control method based on motion capture, including: Acquiring whole body posture data of a target operator acquired by a motion capture device; Converting the whole body attitude data into a coordinate system to obtain an object target position sequence corresponding to a target moving object coordinate system; redundant motion detection is carried out on the object target position sequence, and adaptive smoothing is carried out on the object target position sequence according to the result of the redundant motion detection, so that a smooth object control instruction is obtained; and inputting the smooth object control instruction to a driving controller of the target moving object. In a second aspect, the present application provides a motion object control system based on motion capture, comprising: The acquisition module is used for acquiring the whole body posture data of the target operator acquired by the motion capture device; The processing module is used for converting the whole body posture data into a coordinate system to obtain an object target position sequence corresponding to a target moving object coordinate system; the processing module is used for carrying out redundant motion detection on the object target position sequence, and carrying out self-adaptive smoothing processing on the object target position sequence according to the result of the redundant motion detection to obtain a smooth object control instruction; And the control module is used for inputting the smooth object control instruction to the driving controller of the target moving object. According to the motion capture-based moving object control method and system, the whole body posture data of the target operator acquired by the motion capture device are acquired, then the coordinate system conversion is carried out on the whole body posture data, the joint target position sequence corresponding to the joint coordinate system of the target robot is obtained, redundant motion detection is carried out on the joint target position sequence, self-adaptive smoothing processing is carried out on the joint target position sequence according to the result of the redundant motion detection, so that a smooth joint control instruction is obtained, the smooth joint control instruction is input to the driving controller of the joint of the target robot, and unnecessary joint driving actions and energy consumption are reduced when the joint of the target robot executes capture actions are controlled. Drawings The accompanying drawings, which are incorporated in and constitute a part