Robust Time-Varying Parameter Estimation of Uncertain LPV Systems Subject to Estimated States

In recent years, many techniques for parameter estimation have been developed, with special attention given to the case of time-varying parameters, as they can describe a wide range of physical systems. However, a common requirement is the full-availability of the system states, a condition that may not be fulfilled in some applications. This paper proposes a robust parameter estimation technique for Linear Parameter-Varying (LPV) systems, considering the non-availability of all states and the influence of bounded disturbances. A robust filter with the generalizedH2norm is designed to reconstruct the state vector, based on Linear Matrix Inequalities (LMIs), while a robust adaptive law is formulated to address parameter computation. The convergence of the procedure is proven to be globally uniformly ultimately bounded (GUUB). Numerical experiments are conducted with meta-heuristic optimization to tune the hyperparameters of the adaptive law, highlighting the effectiveness of the proposed methodology in estimating the system parameters.