Anti-windup compensator for nonlinear model predictive control

Actuators in control systems have several limitations in mechanically and electrically, and these limitations cause instability and degradation of a controlled performance. In this study, we treat an anti-windup problem of tracking controller for nonlinear 2-link manipulator using a nonlinear model predictive control (NMPC). Proposed method is designed with focusing to a relationship between maximum values of control inputs and a time-constant of a reference trajectory, which is one of tuning parameters of NMPC. In order to achieve a tracking control of a 2-link manipulator with keeping the control input under a limitation of it, the time constant of a reference trajectory in proposed new NMPC is tuned based on limitation of control inputs. At first, we analyze the relationship between time constant of the reference trajectory and the maximum values of the control input by numerical simulations using MATLAB/Simulink. And next, an algorithm to tune the time-constant of the reference trajectory according to input limitation is designed based on analysis results. In order to show an effectiveness of proposed anti-windup compensator, numerical simulations of tracking control of the two-link nonlinear manipulator with input limitations are performed.