Topological Transformation and Neural Network Backstepping Control for Unknown Non-linear Systems

Abstract

For second-order electromechanical systems with state coupling and nonlinear disturbances, this paper proposes a predefined-time backstepping control method combining topological transformation and RBF neural networks. Through invertible topological mapping, the original system is transformed into an nth-order normalized equivalent form. A predefined-time control framework is introduced using backstepping design with coordinate transformation, guaranteeing convergence within a user-specified time.