Based on ELM forged neural control for a class of strict feedback stochastic nonlinear switched system with time varying delay

In this paper, the stochastic neural network stabilization problem for a class of switched systems with Itô stochastic nonlinear subsystems is investigated. Based on the energy attenuation level of subsystem, backstepping method and neural network, a forged neural networks switching control scheme is designed so that the considered switched dynamic system is globally asymptotically stable in probability. In the designing process of such control scheme, only a single-hidden layer feed-forward neural network (SLFNN) is employed to approximate nonlinear part of switched dynamic systems. Different from the existing neural networks control scheme, the control law only depends on the system state instead of the parameters of neural network; the parameters of the SLFNN are adjusted based on extreme learning machine (ELM). Finally, the proposed control scheme is applied to an example and the simulation results demonstrate good performance.