Autonomous Motion of a Mobile Robot based on Potential Fields and Polar Control

Abstract

Autonomous motion of mobile robots is an open problem in robotics. Challenges in this regard involve the proper interpretation of the information coming from the sensors, and the adequate motion of the robot based on that information to reach a goal without collisions. In this work, we propose a framework that smoothly drives a mobile robot through a collision-free trajectory. The generation of trajectories is based on motion planning using Artificial Potential Fields and the sensed depth information from the environment. The generated path is then followed by an iterative closed-loop feedback controller based on polar coordinates which is guided by the potential field. With this framework the robot can autonomously move to a desired goal avoiding obstacles online. The framework continuously plans its trajectory, being able to avoid obstacles online. Results were obtained using a dynamic simulator and a differential-drive mobile robot that uses an onboard Lidar.