New Initialization Method of Basic Probability Assignment and Application in Cross-Country Environment Perception

Autonomous navigation in cross-country environments presents many new challenges including obstacle perception for Unmanned Ground Vehicle. This paper proposes new method that is suitable for distributing Basic Probability Assignment (BPA), based on which D-S theory of evidence is employed to integrate sensors information and recognize the obstacle. Firstly, select the experiment formula, then by using the fuzzy interpolation we obtain the subordinate to each event in identification framework, secondly, assume the subordinate be equal to correlation coefficient in the formula, which means each feature evidence corresponding to every element. The last step is to extract feature evidence and calculate the Basic Probability Assignment based on the above, and the current obstacle is identified by applying Dempster´s combination rules. The test results show the validity of this new method and its application in cross-country environment perception.