基于分层多模自适应滤波算法的无人机多重传感器故障诊断

应用标准的多模自适应滤波算法能够在较短的时间内检测出系统的单一故障,但是当把它用于检测系统的双重或多重故障时,这一算法需要建立所有可能出现的故障模型,而每一个模型都要对应一个卡尔曼滤波器,需要大量的滤波器并行运算,大大增加了系统的故障诊断时间,为了简化算法并减少算法计算时间,本文提出了一种用于复杂系统的多重故障诊断的分层多重模型滤波技术,在确定某一单个故障发生后,则可以启用一组基于上一单个故障的新滤波器来检测系统的第二重故障,这样减少了并行运算的滤波器数量,从而减少计算量和故障诊断时间.本文将此算法应用于某无人机多重传感器的故障诊断,仿真结果验证了该方法的有效性.

The Hierarchical Multiple Model Adaptive Filter Algorithm Applied to Multiple Failure Detection and Identification of sensors for UAV

The ordinary multiple model adaptive filter algorithm can only detect single fault.When it is used to detect dual and multiple faults,a number of filters are needed in computation.So it is a costly algorithm.A hierarchical multiple model adaptive filter algorithm is presented to detect dual and multiple faults of complex system.In this way,the computation cost and fault detection time is decreased largely.Simulation examples validate the effectiveness.