Sonar-based collision avoidance and probabilistic motion model for mobile robot navigation

This paper addresses a closed-loop control for a mobile robot in navigation. The sonar-based obstacle avoidance algorithm utilizes a region partition strategy by analyzing three forward sensory regions and finds out a navigable region candidate. A probabilistic motion model and the Monte Carlo method are used to estimate the robot pose. The supervisor inside the controller makes a decision to allow the robot to navigate in a cluttered environment successfully. Moreover, experiment results to validate the proposed control scheme in navigation are also presented.