Antidisturbance Distributed Lyapunov-Based Model Predictive Control for Quadruped Robot Formation Tracking

Abstract

This article proposes a novel distributed Lyapunov-based model predictive control (DLMPC) algorithm for a team of autonomous quadruped robots (AQRs) to achieve the formation tracking. We describe the motion characteristic of the AQR members in a disturbed environment with a limited number of inputs. Within the local framework, an anti-disturbance strategy is designed to improve the control performance, where an observer is presented in both state prediction and auxiliary controller design. We explicitly consider the AQR’s dynamic balance and speed limits to enhance the trajectory feasibility. By predicting the neighbors’ behavior, a penalty for inter-AQR collision is introduced with regard to a buffer zone. We analytically address the closed-loop robust stability by the design of a worst-case contraction. Hardware experiments with the real-world distributed control system witness the superior real-time control performance and robustness of our algorithm.