This chapter elucidates the modern theory of finite impulse response (FIR) prediction and Receding Horizon (RH) FIR filtering. It also considers some of the most interesting RH FIR solutions and notice that the zero time step makes the RH FIR and FIR estimators equal in continuous time. Since the FIR predictor and the RH FIR filter can be transformed into each other by changing the time index, it mainly focuses on the FIR predictors due to their simpler forms. There are also ways to find recursive forms for the error covariance of the Unbiased finite impulse response predictor. Suboptimal RH FIR filters subject to the unbiasedness constraint were originally obtained for stationary stochastic processes and for nonstationary stochastic processes. When state feedback control is required for nonlinear systems, then linear estimators generally cannot serve, and extended predictive filtering or prediction is used.
FIR Prediction and Receding Horizon Filtering
Optimal and Robust State Estimation ; 251-273
2022-08-09
23 pages
Article/Chapter (Book)
Electronic Resource
English
Modified Receding Horizon Guidance Law
British Library Conference Proceedings | 1999
|