Parameter Convergence for Adaptive Control in Nonlinear System
S. Shadab, J. Hozefa, Sushama Wagh, N. M. Singh
Abstract
Feedback linearization uses an effective linearizing control to make the input-output dynamics of a nonlinear plant linear such that various linear control strategies can be applied for tracking the output trajectories to the desired one. One of the key open issues in adaptive control of feedback linearizable structures is the estimation of the coefficients of the feedback control law. Therefore, a complete structure is explored without any assumptions or prior knowledge of the system parameters using the finite-time parameter estimator. The weights of the system dynamics are estimated and updated online to robustify the exact cancellation of non-linear terms required by the linearization method. The proposed method exhibits the improved transient response and global convergence of coefficients. It can be applied as a general tool for developing a high-quality control system for a higher dimensional system which is exactly input-output linearizable with uncertain plant parameters. The simulation results of inverted pendulum and buck converter has verified the efficacy of the proposed methodology.