弹道学报2025，Vol.37Issue(4)：57-66,10.DOI:10.12115/ddxb.2024.11009
基于多核融合RBF神经网络的高超声速飞行器自适应滑模控制
Adaptive Sliding Mode Control of Hypersonic Vehicle Based on a Multi-kernel Fused RBF Neural Network
摘要
Abstract
To address the control challenges of hypersonic vehicle arising from unknown complex nonlinearities in the model,an adaptive sliding mode control method based on a multi-kernel fused radial basis function(RBF)neural network was proposed.By constructing a multi-kernel fused radial basis function neural network,the complex nonlinear functions in the hypersonic vehicle model were accurately estimated.Based on adaptive theory,a real-time update strategy for the neural network weights was designed to ensure estimation convergence and accuracy.Furthermore,the hypersonic vehicle control system was decomposed into velocity and altitude subsystems.The altitude subsystem,which contains more state variables,was simplified using backstepping theory.According to sliding mode control theory,control laws for both subsystems were designed,and the asymptotic convergence of the designed controllers was proven based on Lyapunov stability theory,ensuring the stability of the control strategy.Simulation results show that under complex disturbance conditions,the proposed method enables accurate tracking of key state parameters such as velocity and altitude,outperforming traditional methods and significantly improving both stability and control precision of the hypersonic vehicle control system.关键词
高超声速飞行器/径向基函数神经网络/滑模控制/自适应理论/控制律Key words
hypersonic vehicle/radial basis function neural network/sliding mode control/adaptive theory/control law分类
航空航天引用本文复制引用
戴一笑,常思江,陈琦..基于多核融合RBF神经网络的高超声速飞行器自适应滑模控制[J].弹道学报,2025,37(4):57-66,10.基金项目
国家自然科学基金(52202475) （52202475）
