MBC: Multi-Brain Collaborative Control for Quadruped Robots

TL;DR

A MBC:Multi-Brain collaborative system that incorporates the concepts of Multi-Agent Reinforcement Learning and introduces collaboration between the Blind Policy and the Perceptive Policy improves the robot's passability and robustness against perception failures in complex environments, validating the effectiveness of multi-policy collaboration in enhancing robotic motion performance.

Abstract

In the field of locomotion task of quadruped robots, Blind Policy and Perceptive Policy each have their own advantages and limitations. The Blind Policy relies on preset sensor information and algorithms, suitable for known and structured environments, but it lacks adaptability in complex or unknown environments. The Perceptive Policy uses visual sensors to obtain detailed environmental information, allowing it to adapt to complex terrains, but its effectiveness is limited under occluded conditions, especially when perception fails. Unlike the Blind Policy, the Perceptive Policy is not as robust under these conditions. To address these challenges, we propose a MBC:Multi-Brain collaborative system that incorporates the concepts of Multi-Agent Reinforcement Learning and introduces collaboration between the Blind Policy and the Perceptive Policy. By applying this multi-policy collaborative model to a quadruped robot, the robot can maintain stable locomotion even when the perceptual system is impaired or observational data is incomplete. Our simulations and real-world experiments demonstrate that this system significantly improves the robot's passability and robustness against perception failures in complex environments, validating the effectiveness of multi-policy collaboration in enhancing robotic motion performance.

Figures (9)

• Figure 1: We conducted a long-distance test on our controller. At the beginning of the map, the robot relied on height-map and proprioception to traverse through terrain. During the test, we simulated a scenario where the lidar suddenly malfunctioned (by covering it with a orange bag). The robot did not experience any mode crashes and was still able to handle complex terrains effectively.
• Figure 2: Two-stage multi-brain game collaborative training overview.
• Figure 3: Regularization determines whether the terrain is familiar by reconstructing the elevation map, and guides and balances the Blind Policy and Perceptive Policy
• Figure 4: Robustness testing In simulation, the perception-based RMA mode collapses when the height map is corrupted while our policy works well.
• Figure 5: Climbing a wooden box with Lidar