一种全局最优的多基阵声纳数据融合方法

优化分布式检测系统性能的方法和算法很多，但是，这些方法或是优化传感器的判决规则或是优化融合中心的融合规则，而不是两同时优化。本提出了一种同时优化传感器判决 融合中心融合规则的全局优化算法，它扫描所有可能的解，找出在Ｎｅｙｍａｎ－Ｐｅａｒｓｏｎ意义上的最优解，该算法应用于具有三部声纳的分布式检测系统，获得了满意的检测性能证明了算法的有效性。

An Algorithm for Global Optimization of Distributed Sonar Data Fusion

Much research has been done recently in distributed detection system with multiple sensors. But existing algorithms for optimizing system performance focus only on one aspect: either on sensor level or on fusion center. We present an algorithm for global optimization. Fig.1 gives the schematic diagram of a distributed detection system with multiple sensors. After a lengthy derivation, we get Eqs.(7) and (8) for optimization at sensor level, and Eqs.(16) for optimization at fusion center. In section 2, we give five steps for global optimization which simultaneously optimizes the system performance at both the sensor level and the fusion center. The algorithm scans all possible solutions and picks out the optimal one in the Neyman Pearson sense. It was applied to distributed detection system with three sonars, and satisfactory detection performance(Fig.2) is obtained.