基于D-S证据理论的多传感器多特征目标识别

 传统多传感器环境下的目标识别方法主要有两种：利用多传感器获得的数据进行数据融合、利用每个传感器信号的特征向量进行特征融合。但这两种方法均存在目标识别精度不高的问题。针对这一问题，本文提出了一种基于D-S 证据理论两次组合规则的融合方法。该方法在提出多传感器目标识别系统模型的基础上，运用D-S 证据理论对单传感器的多特征信息进行数据融合；根据传感器接收信号信噪比来确定传感器可信度，将该可信度作为D-S 证据理论组合规则中的证据权值，以此来完成目标识别。本文提出的方法综合考虑了传感器的多特征信息和传感器的可信度，克服了传统的D-S 证据理论对证据冲突处理能力有限的缺陷。实验结果表明，该方法具有较高的正确性和有效性，提高了目标识别的精度。

Target Identification Using Multiple Sensors Based on D-S Evidence Theory Characteristics

The traditional target recognition based on multi-sensor environment has two methods. On the one hand, the data obtained from the multi-sensor are used for data fusion, on the other hand the signal obtained from each sensor is used for feature fusion. The two methods both have the problem that the target recognition accuracy is not high. In order to solve this problem, this paper presents a fusion method based on D-S evidence theory. On the basis of a multi-sensor target recognition system model, D-S evidence theory is used for data fusion based on the multi-feature information of a single sensor. According to the sensor signal-noise ratios of the received signal the credibility is determined, which is taken as the weight of evidence of the weighted combination of D-S evidence theory rule to complete the target recognition. This method considers many characteristics information of the sensor and the reliability of the sensor, overcomes the defect that evidence conflict management ability is limited by the traditional D-S evidence theory. Experimental results show the correctness and validity of this method, as well as the improved accuracy of target recognition.