雷达目标识别中获取基本概率赋值的方法

Study on the Method of Constructing basic Probability Assignment Function in Targets Identification

由于传感器的精度、系统组成的许多环节、外部环境影响以及数据的后处理等因素的影响,雷达目标识别中存在大量的不确定性。在这种情况下,利用精确的方法处理不确定性信息,必然造成识别误差的提高。证据理论是近年发展起来的一种新型推理技术,由于它能够处理由不知道引起的不确定性,因此被广泛地应用于目标识别中。然而,在利用证据理论时,基本概率赋值的构建是证据理论实际应用中面临的最大难题。针对雷达目标识别具体问题,在系统分析证据理论的基础上,提出构建基本概率赋值的方法,并通过仿真实验验证了方法的可行性。

Because of the precision of the sensors, the influence of the environment and the processing of the data, there are a lot of uncertainty. If we use the accurate method to handle the uncertainty, the recognition rate will degrade. D-S theory is a new inferential technology, it can solve information uncertainty and unknown in targets identification, but constructing basic probability assignment function is the most difficult problem in practical. In this paper,a method of constructing basic probability assignment function is presented based on analyzing the theory of D-S evidence, and experimental results show that this method is feasible.