Energy Efficient and Conflict-Free Path Planning for Autonomous Mobile Robot Fleets via Obstacle Persistence Evaluation and Smart Zoning

The integration of Autonomous Mobile Robots (AMRs) into smart manufacturing systems offers promising gains in flexibility and efficiency. However, navigating dynamic and constrained industrial environments remains challenging. AMRs often need to brake or stop abruptly to avoid collisions caused by unexpected changes or unforeseen nearby obstacles, which leads to high energy consumption and reduced operational efficiency. Unlike traditional methods that solely focus on safety, using priority rules or speed limits to avoid multi-robot deadlocks, our approach also considers energy efficiency and enables smooth navigation without imposing hierarchy among a fleet of robots. A hybrid framework is introduced that enables energy-aware and conflict-free path planning, enhancing long-term autonomy in AMR fleets. First, the decentralized AMR fleet operates autonomously while continuously reporting potential collision locations, along with the corresponding Time-To-Collision (TTC) values for unforeseen obstacles. Then, the data collected from the fleet is consolidated by a map coordinator unit, which introduces a novel obstacle persistence evaluation method. This method leverages Exponentially Weighted Moving Average (EWMA) modeling to distinguish between temporary and persistent obstacles. Next, to prevent multi-robot deadlocks, we propose a smart zoning framework based on energy and reachability analysis. Finally, an adaptive path planning algorithm is employed, using the updated multi-layered costmap to balance between path length and conflict cost, adapting to spatiotemporal variations in high-traffic locations. Through comparisons with several state-of-the-art algorithms, we demonstrate that the proposed method ensures safe navigation while achieving smoother trajectory execution and effective energy optimization in dynamic environments.