Escaping Local Minima: Hybrid Artificial Potential Field with Wall-Follower for Decentralized Multi-Robot Navigation

TL;DR

By incorporating a wall-following (WF) behavior into the APF approach, this method allows robots to escape local minima, even in the presence of nonconvex and dynamic obstacles including other robots.

Abstract

We tackle the challenges of decentralized multi-robot navigation in environments with nonconvex obstacles, where complete environmental knowledge is unavailable. While reactive methods like Artificial Potential Field (APF) offer simplicity and efficiency, they suffer from local minima, causing robots to become trapped due to their lack of global environmental awareness. Other existing solutions either rely on inter-robot communication, are limited to single-robot scenarios, or struggle to overcome nonconvex obstacles effectively. Our proposed methods enable collision-free navigation using only local sensor and state information without a map. By incorporating a wall-following (WF) behavior into the APF approach, our method allows robots to escape local minima, even in the presence of nonconvex and dynamic obstacles including other robots. We introduce two algorithms for switching between APF and WF: a rule-based system and an encoder network trained on expert demonstrations. Experimental results show that our approach achieves substantially higher success rates compared to state-of-the-art methods, highlighting its ability to overcome the limitations of local minima in complex environments

Figures (7)

• Figure 1: Challenges in decentralized navigation with limited environmental information. The problematic paths of robots due to incomplete environmental data (left) which can be overcame using our proposed method (right).
• Figure 2: Overview of the proposed system. RS determines the rotation angle $\theta_{\text{rot}}$ based on the observation, controlling the navigation mode. If LS is used together, it overrides the RS output for more efficient navigation learned from human demonstrations.
• Figure 3: Four environments for data collection. Three colored robots can be intervened by a human expert while white robots perform as dynamic obstacles. Initial robot positions and goals are randomly generated.
• Figure 4: Test environments where robots are in white, goals in pink, and static obstacles in purple. (a) Real-world layouts with Nakwon (left) and Sogang (right) where robots and goals are placed at random. (b) Symmetric layouts with Flat, Cylind, and Swap from the left.
• Figure 5: Success and arrival rates of real-world layouts