A model-free and finite-time active disturbance rejection control method with parameter optimization

Abstract

In the field of control, although active disturbance rejection control does not rely on the precise system models, it has not achieved completely model-free control. Moreover, this method also faces challenges such as complex structure and difficult parameter tuning. In view of this, a novel model-free and finite-time active disturbance rejection control method based on parameter optimization and filter is proposed in this paper. First, an improved second-order linear extended state observer is proposed based on the tracking error. The proposed observer can not only achieve complete model-free operation and a concise construction, but also synergistically improve the system tracking and estimation performance. Second, a feedback control law is presented based on the outputs of the proposed observer and the specifically designed filter. This control law reduces the computational complexity and avoids the high-frequency chattering phenomenon of the error-feedback control law based on transient process. Third, the system controller is constructed by compensating for the disturbance estimated by the proposed observer in the designed feedback control law. Following that, the finite-time convergence of the proposed observer and the system tracking error under the proposed controller is proven based on the Lyapunov stability theory. Fourth, the parameters of the proposed control method are tuned based on particle swarm optimization algorithm with the specifically designed objective function. Compared with the traditional trial-and-error method, this optimization strategy improves the efficiency and effectiveness of parameter tuning. Finally, simulation experiments have been carried on to compare the control performance among the proposed method and its four variants, as well as four state-of-art controllers. Also, the effectiveness and superiority of the newly-designed strategies are further verified.