Research on Neural Adaptive Control of Quadrotor

Research on Neural Adaptive Control of Quadrotor

연구 분야: Software Development

학회: 2025 37th Chinese Control and Decision Conference (CCDC)

This paper focuses on the strong uncertainty and external interference of quadrotor during rapid flight. Based on the backstepping design, adaptive control methods are adopted, supplemented by neural network intelligent means, to conduct in-depth research on high-performance flight control algorithms, providing ideas for the design of control algorithms for underactuated systems. The main content includes the following: Firstly, in response to the problem of steady-state errors in the nominal backstepping controller, integral backstepping control technology is introduced based on the nominal backstepping method to design integral backstepping control; Afterwards, adaptive control technology is introduced to address the existing mathematical modeling errors, enabling online estimation of important inertial and aerodynamic related uncertain parameters; Finally, a neural network backstepping control is designed to address the unmodeled model in the model, and to perform online approximation of the complex nonlinearity ignored during the modeling of quadrotor.