一种动态多传感器一致性融合算法

针对现有一致性融合算法在处理时变系统的状态估计时,不能准确度量传感器的一致性和可靠性,且传感器一致性均值和可靠性度量存在＂数据饱和＂和＂历史信息浪费＂等问题,将一致性均值和方差的计算转化为时变参数估计问题,引入一致性衰减因子和方差衰减因子,更为客观地度量传感器的一致性和可靠性,实现传感器融合权重的动态调整,从而将一致性融合算法推广应用到时变系统。仿真结果表明,该方法可更为合理地分配各传感器的融合权重,改善一致性融合算法的性能。

A Consensus Data Fusion Algorithm for Dynamic Multi-sensors

Existing consensus fusion algorithms can not accurately measure the consensus and reliability of sensor in state estimation of time-varying system, and ‘data saturation＇ and ‘waste of history information＇ exist in measurement of sensor′s consensus mean value and reliability. In this paper, the calculation of consensus mean value and variance is transformed into the estimation of time-varying parameters, the consensus attenuation factors and variance attenuation factors are employed to measure the consensus and reliability of sensor more objectively, and the dynamic adjustment of sensor fusion weight is realized. Simulation results show that the present approach can reasonably distribute the fusion weights of sensors to optimize the fusion algorithm.