A Distributed Multi-Robot Path Planning Algorithm for Searching Multiple Hidden Targets

Abstract

This paper proposes a new distributed multi-robot path planning algorithm based on fuzzy logic control and reinforcement learning, which can navigate for robots searching for hidden targets in unknown environment. The algorithm is composed of two controllers, a fuzzy logic controller based on multiple behavior coordination strategy and a policy controller based on deep reinforcement learning. The first controller is used for roaming search to find the hidden targets, and the role of the second controller is to navigate from the current position to the targets. To check the performance of the algorithm, we simulate it in simulation environment built in the CoppeliaSim simulation software and implemented by e-Puck robots. The simulation results demonstrate the effectiveness of the proposed method.