Data-driven trajectory tracking control for UAVs

In this study, we confront the issue of advancing the control systems of unmanned aerial vehicles (UAVs) amid denial-of-service (DoS) attacks. Our approach introduces an innovative control schema that leverages a refined dynamic linearized data model. We present a novel technique for designing continuous functional dynamic linearization (CFDL) data models, which facilitates the creation of an adaptive control algorithm devoid of a model dependency—the data-driven model-free adaptive control (DMFAC)—to robustly counteract the adverse effects of DoS attacks. In order to enhance the proposed control model, the least squares method has been utilized to formulate a criterion function for optimal predictive parameter dynamic (PPD) estimation, which is achieved through the reduction of the error sum of squares. This method augments the precision of estimating parameters, elevates the tracking efficacy for parameters with temporal variations, and strengthens the resilience of the control mechanism. The robustness of the CFDL-MFAC algorithm, which is grounded in learning principles, has been meticulously examined, and simulation studies provide evidence to confirm its efficiency.