Multi-Robot Online Coverage Path Planning: Enhanced Collaboration and Task Allocation Strategies for Area and Target Coverage Tasks
Zikai Wang · 2025
Multi-robot online coverage path planning is essential for applications such as environmental monitoring, search and rescue, and urban surveillance. However, it faces significant challenges in dynamic and unknown environments. Key issues include enhancing collaboration among multiple robots for coverage tasks, reducing conflict risks during these tasks, and improving the performance in online target coverage scenarios. This thesis addresses these challenges through a comprehensive exploration of multi-robot online coverage path planning, with a particular emphasis on higher-level algorithmic design rather than local path planning. By focusing on overarching strategies and collaboration frameworks, our approach seeks to optimize the robots’ collective behavior and adaptability in unpredictable settings. To enhance collaboration among robots, our thesis introduces the Artificial Potential Field Based Multi-Robot Online Coverage Path Planning approach (APF-CPP). This innovative method utilizes artificial potential fields to develop individualized coverage strategies for each robot. By fostering better coordination, APF-CPP allows for adaptive task allocation in response to real-time environmental changes. This approach significantly improves coverage efficiency, ensuring that robots can work together seamlessly while adapting to dynamic conditions. To tackle the issue of conflict risks among robots during coverage tasks, we present MAC-Planner, a unified framework that integrates Multi-Robot Task Allocation with Coverage Path Planning. This framework dynamically assigns tasks and plans coverage paths based on real-time updates of task completion status, effectively minimizing potential conflicts. By reformulating area coverage into a point coverage problem, MAC-Planner employs K-means clustering and pairwise optimization to ensure equitable task distribution among robots, fostering cooperation and reducing the likelihood of overlapping paths. Finally, to enhance the performance of multi-robot systems in online target coverage tasks, we introduce the Delaunay-graph-based Multi-Robot Online Target Coverage Path Planner (DMT-CPP). This approach maintains a Delaunay graph in real time for efficient task allocation, utilizing K-means clustering and pairwise optimization for equitable task distribution. Extensive experiments validate its exceptional coverage efficiency and robustness compared to state-of-the-art methods.