Data-Driven trajectory tracking control for non-imitating low altitude vehicles

Aiming at the problems of environmental uncertainty, unknown model parameters and highly nonlinear system dynamics faced by the nonlinear multi-agent system (MAS) in complex, dynamic environments, an iterative model-free multi-agent control algorithm is adopted to construct a control strategy on-line through the real-time collected input/output data, to circumvent the modeling error and enhance the system robustness against unknown modeling errors and enhance the robustness of the system to unknown perturbations. It is concluded that the method can improve the control efficacy and adaptability of the system. The main contributions of this paper are: 1) the construction of a local linearized model by dynamic linearization technique, which reduces the dependence on the system model; 2) the proposal of an adaptive pseudo-partial derivative (PPD) estimation mechanism, which improves the tracking accuracy and enhances the adaptability of the algorithm in time-varying scenarios.