基于联邦滤波的多传感器主动容错估计方法

摘要

针对冗余传感器系统的容错问题，改进一种有反馈式联邦滤波的容错估计算法。通过残差法对硬故障进行检测，并隔离故障子系统的估计值。若全部子系统出现硬故障，取上一时刻的联邦滤波的最优值，通过系统状态方程计算得到系统状态量的预测值进行容错补偿。定义可信度因子，设计可信度单调性法则检测子系统软故障，并设计校正策略。大量仿真实验表明，改进后的反馈式联邦滤波的容错估计算法，可在保证系统估计精度的前提下，提高多传感器系统的容错性与鲁棒性。%A fault-tolerant algorithm based on feedback federated filter is proposed for the redundancy sensors fault tolerant prob-lem.Firstly,the hard fault is detected by using residual error method,and the evaluation value of faulted sub-system is isolated. If all the sub-system are detected as fault,then the optimal evaluation value from last time is used to take place of the fault value, and the state prediction is computed by system state equations to achieve fault-tolerant compensation.Secondly,the reliability factor and the reliability monotonicity rules are designed to detect soft fault.In the experiments,it is shown that the proposed model not only ensures the system accuracy,but also improves the multi-sensor system fault-tolerant ability and robust perform-ance.