Adaptive Neural Network-Based Nonlinear Control of Quadcopters

DOI: https://doi.org/10.55228/JTST140411

Keywords

Quadcopter, Adaptive Control, RBF Neural Network, Estimating Disturbance

Abstract

This paper presents a novel adaptive neural network control strategy for trajectory tracking control of quadcopters in the presence of external disturbances and model uncertainties. The proposed method utilizes a hierarchical control structure, where an outer position loop is structured from a basic sliding mode control (SMC), and the inner-loop attitude control is comprised of a backstepping approach and an adaptive Radial Basis Function (RBF) neural network. The RBF neural network is designed to approximate lumped disturbances in real time through Gaussian basis functions and an online weight adaptation law, eliminating the need for detailed disturbance modeling. To evaluate the performance of the proposed approach, we conduct comparative simulations against a SMC controller and a Robust Feedback Linearization (RFBL) control method. Results obtained demonstrate that the RBF-based controller achieves superior tracking accuracy, faster convergence, and improved disturbance rejection, particularly under time-varying and uncertain conditions. These findings highlight the potential of adaptive learning-based controllers for robust and model-free UAV applications.

Downloads

Download data is not yet available.