Design Of Robust LQR Control For Nonlinear System Using Adaptive Dynamic Programming
Shantakumari Rajan, Rajan Shalini P · Int. J. of Aquatic Science · 2021
One of the important challenges in the design of LQR for real time applicationsis the optimal choice state and input weighting matrices (Q and R), which play a vital rolein determining the performance and optimality of the controller. Commonly, trial anderror approach is employed for selecting the weighting matrices, which not only burdensthe design but also results in non-optimal response. Hence, to choose the elements of Qand R matrices optimally, Adaptive Dynamic Programming (ADP) algorithm is used forselecting the most suitable Q and R matrices by iteration which reduces the performanceindex of the system to be considered. However, stability is only a bare minimumrequirement in a system design. Ensuring optimality guarantees the stability of thenonlinear system. Dynamic programming is a very useful tool in solving optimization andoptimal control problems by employing the principle of optimality. There are severalspectrums about the dynamic programming. One can consider discrete-time systems orcontinuous-time systems, linear systems or nonlinear systems, time-invariant systems ortime-varying systems, deterministic systems or stochastic systems. The inverted pendulumis a standard benchmark control problem and for the control of which numerous controlalgorithms have evolved over the ages. The main objective of this project is to design arobust linear quadratic regulator (RLQR) for nonlinear system using adaptive dynamicprogramming and to propose an optimal tracking control approach based on adaptivedynamic programming (ADP) algorithm in order to solve the linear quadratic regulationproblems for nonlinear systems in an online fashion.