Passivity and Immersion based-modified gradient estimator: A control perspective in parameter estimation

In this paper, a constructive and systematic strategy with more apparent degrees of freedom to achieve the accurate estimation of unknown parameters via a control perspective is proposed. By adding a virtual control in the final equation of the gradient dynamics, the Gradient Estimator (GE) and Memory Regressor and Extension (MRE) approaches are extended. The solution of the virtual control law is identified by the P&I approach. The P&I approach is based on the choice of an appropriate implicit manifold and the generation of a suitable passive output and a related storage function. This facilitates the virtual control law being obtained in a way that the parametric error converges asymptotically to zero. Because the above ideas connect with the P&I approach and GE, the developed methodology is labeled the passivity and immersion-based modified gradient estimator (MGE). The proposed P&I-based modified gradient estimator is extended via the MRE approach. This modification provides improved transient response and fast convergence. Based on certain PE and non-PE examples, a comparative analysis is carried out to show the efficacy of the proposed approaches.