Research on LQ Controller Design and its Parameter Tuning using Error-based Extended State Observer

Although LQ(Linear Quadratic) control is simple in structure and easy to implement, it suffers from the degradation of its control performance for an essential nonlinear plant as this control method is based on a linear model. The ESO(Extended State Observer) estimates not only the model uncertainty but also the external disturbances as total disturbances. We first construct an extended state equation by introducing a new state which contains non-linearity, model uncertainty and disturbance. Since the extended state equation is a linear state equation, the LQ controller design method can be used for this equation. Then, the important parameters of the LQ controller, the state weight coefficient matrix and the input weight coefficient matrix, are determined using the diffusion-type PSO(Particle Swarm Optimization)-gradient hybrid optimization algorithm. The obtained controller is applied to the stabilization control for IWP(Inertia Wheel Pendulum) to verify its effectiveness.