基于状态失配补偿的机器人动作控制方法、系统、设备
By acquiring time delay-related information of multi-source historical state sequences, determining state reliability and assigning weights, and performing compensation encoding and modulation, the problem of multi-source state mismatch in robot control is solved, and the accuracy and stability of motion generation are improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANTONG INST OF TECH
- Filing Date
- 2026-05-18
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, multi-source historical states in robot control suffer from problems such as asynchronous sampling, transmission delay, local frame loss, and inconsistent cross-source descriptions, resulting in inaccurate motion generation and modulation, motion output oscillation, peak overshoot, and cumulative trajectory deviation.
By acquiring the robot's multi-source historical state sequence and its time delay related information, the state reliability description information is determined and weighted, and compensation encoding is performed to generate a state representation aligned with the current control time. This state representation is then used to modulate the action feature sequence, and finally, the robot control action is output.
It improves the consistency of robot control output response to the current real state, reduces action misdirection, and enhances the accuracy and stability of motion control under multi-source asynchronous conditions.
Smart Images
Figure CN122401403A_ABST