Augmented obstacle avoidance controller design for mobile robots

Abstract

For unicycle robots and an arbitrary continuous reference tracking controller, we propose a mildly invasive obstacle avoidance augmentation only activated in an eye-shaped neighborhood around an obstacle and providing avoidance guarantees. Switching between different control phases is orchestrated by a hybrid automaton with hysteresis mechanisms that prevent Zeno behavior. Since the size of the eye-shaped neighborhood and its orientation depend continuously on the orientation and velocity of the robot, the velocity control input is continuous. Numerical simulations illustrate the performance of an avoidance-augmented tracking controller.