Neural Adaptive Sliding Mode Control with Optimized Fast Logarithmic–Tanh Sliding Surface for Quadrotor UAVs under Parametric Uncertainties and External Disturbances

Abstract

This paper presents the design, stability analysis, and validation of a neural adaptive sliding mode controller with an optimized fast logarithmic-tanh sliding surface (NASMC-OFLTS) for the attitude and altitude control of quadrotor unmanned aerial vehicles (UAVs) under parametric uncertainties and external disturbances. A genetic algorithm (GA) is employed to optimally tune the parameters of the proposed fast logarithmic--tanh sliding manifold, thereby enhancing transient response and steady-state tracking accuracy. To compensate for unknown parametric uncertainties, adaptive radial basis function neural network (RBFNN)-based laws are developed for online parameter estimation. The closed-loop stability and convergence of the proposed control scheme are rigorously established using Lyapunov stability theory. The controller is validated through extensive numerical simulations under two challenging scenarios, namely parametric uncertainties and simultaneous parametric uncertainties with external disturbances, and is systematically compared with conventional sliding mode control (SMC) and two neural adaptive sliding mode control schemes (NNASMC-LS and NNASMC-TS). Simulation results demonstrate that the proposed NASMC-OFLTS controller consistently achieves the best tracking performance among all controllers. Compared with the conventional SMC, it achieves improvements of up to 82% in rise time (tr), 79% in settling time (ts), 96% in integral absolute error (IAE), and 69% in root-mean-square error (RMSE), while consistently outperforming both NNASMC-LS and NNASMC-TS across all evaluated performance metrics.