Obstacle Avoidance by DSmT for Mobile Robot in Unknown Environment

Abstract

A method by Dezert-Smarandache theory (DSmT) is proposed for obstacle avoidance of mobile robot in unknown environment. The grid environment map is constructed for robot. On the basis of DSmT, the generalized basic belief assignment (gbba) is defined to evaluate the grid state: empty, has obstacle, and unknown. Then the belief values of the grid state from different time slice are combined by DSmT. Experiments including eleven typical simulation scenes are given. In these experiments, one scene test fails and the rest of ten scenes are successful in which robot can avoid all obstacles. The results show that the method is effective and available for mobile robot's obstacle avoidance in unknown environment.