树形分布式检测系统的反馈加权融合算法

Confidence-weighted Fusion Algorithm with Feedback in Distributed Detection System of Tree Structure

通过分析分布式检测的工作原理,提出了一种基于树形检测系统的带有信心加权的融合算法,采用一种将父节点的判决结果与子节点的判决结果多步反馈比较方法来在线的估计权值,通过给不同的节点赋予不同的权值,从而降低性能差的节点对系统检测性能的影响,分析表明,该算法能有效地提高系统的检测性能。

By analyzing the principle of the distributed detection system,a confidence-weighted fusion algorithm with feedback is proposed in this paper,the weights of the nodes is estimated online by comparing the multi-step feedback decision between the parent nodes and the child nodes,different nodes will have different weights,thus nodes with bad performance have less effect on the detection performance of the system.