A Hybrid Control Strategy for Autonomous Navigation while Avoiding Multiple Obstacles at Unknown Locations

Abstract

We consider the problem of steering an autonomous robot to a target position in the presence of multiple obstacles. We propose a modification of the potential field method where the robot is repulsed from an obstacle. We use a Lyapunov-based hybrid state-feedback controller that, using logic, executes an algorithm suitable for steering a point-mass robot to the target while avoiding multiple unknown obstacles. Global convergence and stability properties induced by the closed-loop hybrid controller are verified in simulations.