Application of a Modified Joint Unscented Kalman Filter for Parameter Estimation of a Class of Mechanical Systems

The unscented Kalman filter (UKF) is a tool for state and parameter estimation in nonlinear dynamical systems. There are two primary approaches for utilizing UKF in parameter and state estimation. The first approach is known as joint filtering, where both states and parameters are estimated concurrently. The second approach, dual filtering, involves the use of two separate filters to estimate states and parameters independently. Recently, a modified joint Unscented Kalman Filter (MJUKF) has been introduced by the author. This modified approach is applicable to a class of nonlinear systems and offers two significant advantages which are improved estimation accuracy and reduced computational complexity. This study investigates the application of the MJUKF for parameter estimation in two selected mechanical systems. The results demonstrate that the modified filter effectively estimates parameters in mechanical systems revealing its advantages over the standard joint unscented Kalman filtering method in terms of both accuracy and computational complexity.