The Application of Improving Space-time DS Evidence Theory in Distinguishing Vehicle

摘要

In this paper, it takes advantage of evidence theory to fuse the data with multi-sensors and multi-measuring periods. It discusses three kinds of fusion structures: concentrated fusion, distributed fusion without feedback and distributed fusion with feedback. In the application of vehicle type distinguishing, through theoretical analysis and simulation results, the paper gets the conclusion that when the data provided by the sensors is not very accurate (even wrong), the distributed fusion without feedback can get the highest rate of correct result, the distributed fusion with feedback follows and the concentrated fusion is the worst.