Controlador lqr a tempo discreto: análise de erro sob controle discreto

Abstract

This research work deals with the design of an optimal controller for discrete linear systems in the presence of uncertainties. The main objective is to develop a control method that ensures the efficiency of the Linear Quadratic Regulator (LQR) despite parametric variations while simultaneously minimizing control error. To achieve this, a constraint was defined to regulate the difference between the optimal and non-optimal controls, and simulations were conducted to analyze the performance of the approach. Using control gains, the associated trajectories were compared, and cumulative costs were calculated. The results obtained showed that the error between the optimal and non-optimal trajectories decreases asymptotically, while the total cost difference is bounded by a parameter that depends on the time horizon.