Collision/Obstacle Avoidance Dynamic Formation Reconfiguration of High-Order Nonlinear Multi-Agent Systems

Multi-Agent formation involves the motion conflicts of large numbers of physical systems, which leads to significant safety challenges in terms of collision/obstacle avoidance. This paper studies the problem of collision/obstacle avoidance formation tracking for a class of high-order strict-feedback nonlinear multi-agent systems and presents a dynamic formation reconfiguration method. First, a smooth and analytic reference trajectory is constructed, which consists of a convex combination of preset trajectory and delayed target trajectory with its coefficient adaptively adjusted by two barrier functions to avoid collisions. Then, a distributed backstepping control scheme with high-order filtering for barrier functions is proposed. Resultantly, all agents adaptively reconfigure their formation shape within the barrier function's coverage to avoid collisions with each other and obstacles, while quickly forming a pre-defined shape of formation following the target trajectory outside the barrier function's coverage. Simulation results are given to illustrate the effectiveness of the proposed scheme.