Robust control of elevator based on Hamiltonian neural network
A nonlinear robust control strategy based on deep neural network was proposed for the uncertain elevator with random oscillation at the mounting base.First,a deep neural network(HNN)with Hamiltonian mechanics priors was employed to fit the dynamic model of the system.Then,based on this model,a robust stabilization controller for the system was designed by combining an implicit Lyapunov(IL)function.Simulation results showed that the proposed controller(HIL)achieved almost identical control performance to the controller(IL)based on the accurate model.Compared with traditional IL control based on black-box neural network models and error models,the convergence time of the proposed controller was reduced by 25.9%(BIL)and 32.5%(IL-E),respectively.The results demonstrated that the proposed control strategy could accurately represent the nonlinear dynamic characteristics of the system,effectively suppress the uncertain influence of base oscillation,and achieve rapid and accurate robust stabilization of the elevator.
base oscillationimplicit Lyapunov functionrobust controldeep neural networkelevator
NETL
NSTL
万方数据
