LSFR算法在多传感器分布式检测中的优化研究

在多传感器分布式检测系统中,常规融合规则算法要求传感器误差概率已知,且系统中传感器和融合中心同时优化存在一定困难.提出最小二乘融合规则(LSFR)算法,算法不依赖噪声环境稳定性以及传感器的虚警概率与检测概率,融合中心根据各个传感器的硬决策,得到全局的硬决策,并在传感器和融合中心处理达到最优时,获得最佳全局性能.仿真结果表明:对比似然比融合决策算法与Neyman Pearson融合规则(NPFR)算法,LSFR算法全局检测概率显著提高,且在不同数量规模传感器和更多类型的分布式检测系统中具有较好兼容性.

Optimization research on LSFR algorithm for distributed detection in multiple sensors systems

In multiple sensor distributed detection system,conventional fusion rule algorithm need to know the error probabilities of each sensor,and it is difficult to optimize the sensors and the fusion center simultaneously in the system. A least squares fusion rule(LSFR)algorithm is proposed,LSFR algorithm does not rely on any stability of the noise environment and false alarm and detection probabilities of the sensors. Fusion center combines the hard decisions of each sensor to make global hard decision,and acquire the optimal global performance when the processing of the sensors and fusion center are optimal. The simulation results show that,compare to likelihood ratio fusing decision-making algorithm and Neyman Pearson fusion rule (NPFR)algorithm,the global detection probability of LSFR algorithm is significantly improved,and LSFR algorithm has preferable compatibility in the distributed detection system of different scale of sensor and more types.