Path planning of a mobile robot using an improved mixed-method of potential field and wall following

Abstract

The existing Bug algorithms, which are the same as wall-following algorithms, offer good performance in solving local minimum problems caused by potential fields. However, because of the odometer drift that occurs in actual environments, the performance of the paths planned by these algorithms is significantly worse in actual environments than in simulated environments. To address this issue, this article proposes a new Bug algorithm. The proposed algorithm contains a potential field function that is based on the relative velocity, which enables the potential field method to be extended to dynamic scenarios. Using the cumulative changes in the internal and external angles and the reset point of the robot during the wall-following process, the condition for state switching has been redesigned. This improvement not only solves the problem of position estimation deviation caused by odometer noise but also enhances the decision-making ability of the robot. The simulation results demonstrate that the proposed algorithm is simpler and more efficient than existing wall-following algorithms and can realise path planning in an unknown dynamic environment. The experimental results for the Kobuki robot further validate the effectiveness of the proposed algorithm.