An Improved Obstacle Avoidance Method for Robot Based on Constraint of Obstacle Boundary Condition

Abstract

Integrating odometer, IMU and laser radar information, we can obtain the robot's current location, heading, as well as the surrounding environment and obstacle information. Using obstacle information, especially boundary information of obstacles, an improved obstacle avoidance algorithm is proposed on the basis of Vector Field Histogram algorithm (VFH). In this algorithm, the robot's feasible direction is determined by using the boundary of the obstacle relative to the position of the robot. It overcomes the sensitivity of the traditional VFH algorithm to the threshold, and overcomes the problem that the adaptive threshold method proposed in some papers has blindness and higher complexity when looking for a feasible direction. Then, according to the reference driving direction, the angular velocity and linear velocity control law of the robot are proposed. The simulation results and the experimental results on Husky robot prove that the proposed algorithm can make the robot avoid the obstacles safely and efficiently to the target point.