Iterated Square Root Unscented Kalman Filter for state estimation — CSTR model

Abstract

In this study, the use of improved Unscented Kalman Filter algorithm based on iterated measurement updates is proposed in an attempt to estimate the nonlinear and non-Gaussian state variables (the concentration and temperature) of the Continuously Stirred Tank Reactor (CSTR) process. Various conventional and state-of-the-art state estimation techniques are compared based on their estimation performance on this objective. These techniques are the Unscented Kalman Filter (UKF), the Square-Root Unscented Kalman Filter (SRUKF), the Iterated Unscented Kalman Filter (IUKF) and the developed Iterated Square Root Unscented Kalman Filter (ISRUKF). The results of the study indicate that the ISRUKF technique has better convergence properties than the IUKF technique; and both of them can provide improved accuracy over the UKF and SRUKF techniques. Moreover, ISRUKF technique is able to provide accuracy related advantages over other estimation techniques. Since this approach re-linearizes the measurement equation by iterating an approximate maximum a posteriori (MAP) estimate around the updated state, instead of relying on the predicted state.