Neural Network Based Flight Attitude Control Using Robust Sliding Mode

In this article, an adaptive sliding mode control (SMC) strategy is designed for the attitude control of aircraft under the influence of dynamic disturbances. A radial basis function neural network (RBFNN) is integrated with the SMC framework, effectively handling the nonlinear aerodynamic forces and moments. This network adapts through the variation in weights and provides a real-time function approximation in order to improve the robustness of the controller against uncertainites. A sigmoid function is incorporated for smoother control input to reduce the chattering effect which is found commonly in SMC. The control strategy has been formulated using only partial information from the aerodynamic model based on available physical parameters of the plant. Its effectiveness is verified through the use of a quadratic Lyapunov function which helps confirm that closed loop system remains stable and the states stay within bounded limits. The simulations show that the control method works well, as it follows the desired attitude commands closely and keeps performing reliably even when different disturbances are introduced.