Obstacle Avoidance for Robots Based on VFH* Method in Uncertain Environment

Abstract

This paper presents an improved VFH* method for robot obstacle avoidance.Follow the Carrot Algorithm is used to offer the local goal points for VFH*.Because of the direction error of sonar sensor in some situations,VFH* method fails to find a way to the goal even the path actually exists.To solve this problem,a const parameter is invited in this paper.Each obstacle is enlarged in different size according the distance to the robot.Another shortcoming of the VFH* method is resulted from the lack of candidate directions.We add one more candidate direction when a valley is too wide.The simulation shows that the modulations are valid.