APF-based Control for Obstacle Avoidance in Smart Electric Wheelchair Navigation

Abstract

In this paper, we propose an autonomous and real-time navigation module that drives a smart electric wheelchair (SEW) to the desired target, along with its ability to avoid obstacles in a dynamic 3D environment. To avoid obstacles, we use a remote sensing technology (Lidar) to provide real-time mapping of the environment. The generated map is fed to the virtual obstacle avoidance control algorithm. To control the SEW to the target, we use an algorithm based on the artificial potential field (APF) method to generate the desired trajectory and simultaneously guarantee its efficiency and continuity. Our algorithm is specially adapted and based on the idea of virtual forces, the one of attraction generated by reaching the desired target and the one of rejection caused by the detected obstacles. The SEW behaviour is obtained by the sum of the force of attraction and all the repulsion forces at a given position. The simulation results in a corridor with obstacles that confirm the viability of the proposed APF algorithm. It can be used effectively to plan the trajectory of intelligent wheelchairs and can be applied in real-time scenarios.