Cooperative Planning and Following Control for Autonomous Vehicle Formations via Historical Path Updates

This article addresses the cooperative planning and following control problems of autonomous vehicle formations in both linear and triangular configurations. A distributed framework that integrates real-time and historical data for formation following tasks is established by incorporating a collision-free planner and a tracking controller. Two practical driving scenarios, differentiated by the availability of adequate passable areas, are discussed to cope with the obstacle-avoidance demands in complex environments. A segment-wise path updating strategy is proposed to achieve formation obstacle avoidance with insufficient historical information in the following driving framework, which utilizes a local planner based on the rapidly exploring random tree (RRT) star algorithm. Besides, the formation transformation technique is employed in circumstances where the passable area is limited. A velocity following strategy is incorporated to formulate the reference model, and a model predictive controller is designed for historical path tracking. Platform simulations and real-world experiments are conducted to demonstrate the validity and efficacy of our proposed framework in both following driving and obstacle-avoidance tasks.