LMMSE Filtering in Feedback Systems with White Random Modes: Application to Tracking in Clutter

摘要

A generalized state space representation of dynamical systems with randommodes switching according to a white random process is presented. The newformulation includes a term, in the dynamics equation, that depends on the mostrecent linear minimum mean squared error (LMMSE) estimate of the state. Thiscan model the behavior of a feedback control system featuring a stateestimator. The measurement equation is allowed to depend on the previous LMMSEestimate of the state, which can represent the fact that measurements areobtained from a validation window centered about the predicted measurement andnot from the entire surveillance region. The LMMSE filter is derived for theconsidered problem. The approach is demonstrated in the context of targettracking in clutter and is shown to be competitive with several popularnonlinear methods.