Mathematical Model and Adaptive Neural Network Flight Control of Wing-in Ground-Effect Vehicle

Based on the analysis of the initial design tasks for ground-effect vehicles, this paper utilizes the Vortex Lattice Method (VLM) in conjunction with the image method to establish a simplified, partially linearized multi-parameter mathematical model of the aerodynamic characteristics of ground-effect vehicles. Furthermore, considering the presence of various disturbances such as gusts and waves on the sea surface, an Active Disturbance Rejection Controller (ADRC) based on an Adaptive Radial Basis Function Neural Network (ARBFNN) has been designed for ground-effect vehicles. The adaptive RBFNN replaces the traditional Extended State Observer (ESO) in the estimation of disturbances and compensates them into the control inputs. Simulink simulation results demonstrate that this approach effectively addresses the issue of inaccurate disturbance estimation in traditional ADRC due to the fixed parameters of the ESO.