A Self-Tuning Kalman Filter with Input Estimation and Its Application

For discrete-time systems with unknown constant input and unknown noise statistics, using the modem time series analysis method, based on ARMAX innovation model, this paper presents a new self-tuning Kalman filter with input estimation. As an application example, a new self-tuning α-β-γ tracking filter with input estimation is presented, and simulation results show its usefulness.