“…This method can be divided into two categories: the adaptive stochastic modeling and adaptive fading filter [18,30]. The adaptive stochastic modelling transforms state disturbances within a particular time period into the state disturbance at the current time point for system state estimation [3,7,24,33,34,38,41]. However, this method requires that residual vectors or predicted residual vectors at each time point be in the identical type, spatial dimension and distribution, which is difficult to achieve in a highly dynamic environment [18,30].…”