Obstacle avoidance and motion control for mobile robots

Autonomous mobile robot is an intelligent machine that is capable of navigating in an obstacle-cluttered environment without collision and without human intervention. The robot uses information provided by its sensors to build a map using graph or grid representation. In this paper the environment is spatial and temporally varying. To find a safe path among obstacles, a map must be created and updated each time an obstacle is encountered. We present an algorithmic work based on virtual force field concept for obstacle detection and avoidance. As a result a collision free path between a start point and a goal position through a flat 2D environment is obtained. The obstacle avoidance technique generates a sequence of speed and steering commands for a vehicle with two front wheels separately driven using two dc motors. A new cross coupled computer control scheme for the two drive motors of the vehicle wheels is proposed. The continuous motion of the vehicle is executed using this control scheme. The new control scheme accounts for drives mismatch. Extensive simulation results are conducted to illustrate the useful aspects of the proposed integrated system.