An Efficient Closed-Loop Adaptive Controller for a Small-Sized Quadruped Robotic Rat

Abstract

Large quadruped robots have shown potential for a wide range of everyday tasks due to their superior terrain adaptation. However, small-scale quadruped robots have limited payloads and thus cannot carry sufficient sensing and computational resources, which imposes limitations on their environmental adaptability. To address this challenge, we proposed an efficient closed-loop adaptive controller by simplified pose estimation and control strategy that utilizes only inertial measurement unit sensors, which drastically reduces the control computation. Accordingly, we integrated this control system into a small-scale quadruped robot, SQuRo, and conducted a series of experiments to verify the environmental adaptation performance of SQuRo. The results demonstrated that SQuRo has achieved 6 kinds of robust motion: slope stabilization motion, linear tracking motion, autonomous fall recovery motion, uneven terrain walk, slope walk, and obstacle avoidance. This work paves the way for small-scale quadruped robots to autonomously perform tasks in challenging environments.