摘要：This paper proposes a finite memory structure(FMS) filter for state estimation in time-varying systems withmeasurement delay, called the TV-FMS filter. By incorporatingtime-delayed measurements, the proposed TV-FMS filter isobtained by directly solving a weighted least-square costfunction. The proposed TV-FMS filter has inherent propertiessuch as unbiasedness, deadbeat, and robustness. Throughextensive computer simulations for three kinds of discretetime-varying noisy signal models with a measurement delay aswell as an uncertain model parameter, the proposed TV-FMSfilter can be shown to be comparable with the Kalman filter withinfinite memory structure (IMS) for a nominal system andbetter than that for a temporarily uncertain system.
