MODIFIED EXTENDED KALMAN FILTERING AND A REAL-TIME PARALLEL ALGORITHM FOR SYSTEM PARAMETER-IDENTIFICATION

The most popular real-time filtering algorithm for nonlinear systems is perhaps the extended Kaiman filter which will be called EKF for brevity. In this note, a modification of the EKF algorithm, which will be called MEKF for short, is introduced. The modification is achieved by an improved linearization procedure. For this purpose, a parallel computational scheme is recommended, and it has immediate applications to identifying unknown system parameters of time-varying linear stochastic state-space models in real-time. It should be noted that just as the EKF, our MEKF is also ad hoc in the sense that it is a real-time approximation method. Numerical examples with computer simulation are included in this note to demonstrate the effectiveness of this new procedure over the EKF algorithm. © 1990 IEEE