Incorporating Control Barrier Functions in Distributed Model Predictive Control for Multirobot Coordinated Control

Multirobot motion planning and control has been investigated for decades and is still an active research area due to the growing demand for both performance optimality and safety assurance. This article presents an optimization-based method for the coordinated control of multiple robots with optimized control performance and guaranteed collision avoidance. We consider a group of differential drive wheeled robots, and design a distributed model predictive control (DMPC) where the team-level trajectory optimization is decomposed into subproblems solved by individual robots via alternating direction method of multipliers (ADMM). Our DMPC design utilizes a discrete-time control barrier functions method to develop control constraints that provide collision avoidance assurance. Compared to existing ADMM-based DMPC methods with Euclidean distance constraint for collision avoidance, our method ensures collision avoidance with minimal compromise of the optimality with respect to primary control objectives. We validated our method in a multirobot cluster flocking problem. The simulation results show effective coordinated control that achieves improved control performance and safety guarantees.