Approach Based on Improved Interval D-S Theory for Target Identification

Abstract

In traditional multi-sensor based target identification, problems of sensor reliability and identification confidences would generate significant influence on practical implications. Therefore, a new identification algorithm based on Improved Interval Dempster-Shafer Theory (IIDST) is proposed in this paper. Specifically, this algorithm models reliability of sensors and identification outputs from sensors as interval values, and then combines practical interval outputs through interval evidence combination rules. Finally, the capability of the IIDST algorithm is evaluated via theoretical simulations with results demonstrating the effectively of the proposed algorithm.