Model-Free Control Design for Feedback-Linearizable SISO Systems
Data-driven control has gained significant attention in recent years, particularly regarding feedback linearization of nonlinear systems. However, existing approaches face limitations when it comes to implementing them on hardware. The main challenges include the need for very small sampling times, which strain hardware capabilities, and the requirement of an initial open-loop data set, which can be impractical for stabilizing unstable equilibrium points. To address these issues, we propose a two-stage model-free approach that combines a high-gain observer and a dynamic controller. This eliminates the hardware implementation difficulties mentioned earlier. The high-gain observer acts as a robust state estimator, offering superior noise attenuation and lower computational costs, crucial factors for digital hardware implementation. Unlike data-driven methods, our design's stability and performance depend on a tunable software parameter, simplifying digital implementation without overburdening hardware resources. Experimental results on a Twin Rotor system demonstrate the effectiveness of our approach compared to the state-of-the-art data-driven method.