Hybrid Feedback for Autonomous Navigation in Environments with Arbitrary\n Convex Obstacles

Abstract

We develop an autonomous navigation algorithm for a robot operating in\ntwo-dimensional environments cluttered with obstacles having arbitrary convex\nshapes. The proposed navigation approach relies on a hybrid feedback to\nguarantee global asymptotic stabilization of the robot towards a predefined\ntarget location while ensuring the forward invariance of the obstacle-free\nworkspace. The main idea consists in designing an appropriate switching\nstrategy between the move-to-target mode and the obstacle-avoidance mode based\non the proximity of the robot with respect to the nearest obstacle. The\nproposed hybrid controller generates continuous velocity input trajectories\nwhen the robot is initialized away from the boundaries of the unsafe regions.\nFinally, we provide an algorithmic procedure for the sensor-based\nimplementation of the proposed hybrid controller and validate its effectiveness\nthrough some simulation results.\n