一种分布式软决策数据融合系统的性能分析

On the Performance of Distributed Soft Decision Fusion System

在分布式决策融合系统中 ,各传感器仅向融合中心传送一位判决信息。由于该判决信息不能充分反映传感器的观测信息 ,故信息损失较大。本文提出了一种分布式软决策融合方法 ,对多个传感器的判决信息及其可信度进行融合 ,并形成最终的判决。将该方法用于航空探潜系统中多部声呐的信息融合 ,收到了良好的效果。

Distributed hard decision fusion method can not make full use of the information observed by multiple sensors. In order to overcome this shortcoming, a distributed soft decision fusion method is presented, which requires that each sensor transmits both its decision and confidence of that decision to the fusion center. Since the decision of each sensor is obtained by comparing its test statistics to a decision threshold, the confidence in a particular decision can be reflected by the departure of the realization of the test statistics from the decision threshold. Taking these local decisions and confidence information as its own observation, the fusion center implements the Neyman Pearson test and makes the global decision. Using this fusion method in a detection system which contains multiple sonars, a satisfactory detection performance is obtained.