为了控制观测值异常对组合导航精度的影响,提出基于交互多模型的观测数据异常影响控制算法,即建立观测正常和异常的多模型集,各模型间通过马尔可夫模型进行切换,当观测值正常时,正常模型起作用;观测值异常时,故障模型起作用,从而可有效地抑制观测数据异常对滤波精度的影响。
Interacting multiple model algorithm is proposed to control the influence of observation outliers in GPS/INS navigation.Redundant observations are not requested in this method.A multi-model set which includes both normal and exceptional observational models is established.The Markov transition probabilities are used to switch one model to another.When observations are normal,the normal observation model is used chiefly for the others the exceptional observation model works.Consequently,the algorithm can effectively control the influence of observation outliers in GPS/INS navigation.