Distributed Shape Discrete Layers Based Formation Control and Global Path Planning for Multi-Agent Systems

Multi-agent formation control is a key challenge in robotics, particularly for achieving complex tasks in obstacle-rich environments. While distributed self-organization methods based on shape discrete layers offer high scalability and adaptability, these approaches are prone to local minima, which can prevent task completion in complex settings. To address this limitation, this paper proposes a hierarchical control framework that integrates the A* global pathfinding algorithm with a distributed shape discrete layer controller. The A* algorithm first generates a safe, globally optimal reference path that accounts for static obstacles. Subsequently, the distributed controller guides a swarm of agents to track this path, while concurrently handling local collision avoidance and self-organizing into a desired formation. Simulation experiments demonstrate that a swarm of 37 agents can successfully follow the global path, navigate a complex obstacle field, and converge into a predefined "snowflake" formation at the target location. The results validate that the proposed hybrid approach effectively guarantees global task reachability while retaining the robustness and scalability advantages of distributed control.