A method for collaborative control of a cluster of legged robots fusing ground action force perception
By integrating ground force perception and a distributed reinforcement learning framework, the problem of adaptability and collaborative control of legged robots in complex terrain was solved, enabling real-time ground perception and dynamic motion adjustment, and improving the adaptability and stability of the swarm.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA NORTH VEHICLE RES INST
- Filing Date
- 2025-08-04
- Publication Date
- 2026-07-21
AI Technical Summary
Existing legged robots have poor adaptability in complex terrain, unstable dynamic motion control, weak multi-robot swarm collaboration, high computational complexity, communication delay and computing power bottlenecks in traditional centralized control methods, and the inability of sensors to effectively perceive ground information in the field, resulting in unstable motion.
By employing a method that integrates ground force perception, and through the construction of a legged robot swarm state space and multiple reward functions, combined with a distributed actor-critic reinforcement learning framework, pressure sensors are used to perceive ground information and generate motion control commands to achieve swarm collaborative control.
By sensing ground changes in real time, dynamically adjusting motion strategies, improving adaptability, maintaining motion stability, optimizing action strategies, and improving cluster collaboration efficiency, this method solves the problems of computational complexity and insufficient sensor perception in traditional methods.
Smart Images
Figure CN120821273B_ABST