A new control method of nonlinear systems based on impulseresponses of universal learning networks

摘要

A new control method of nonlinear dynamic systems is proposednbased on the impulse responses of universal learning networks (ULNs),nULNs form a superset of neural networks. They consist of a number ofninterconnected nodes where the nodes may have any continuouslyndifferentiable nonlinear functions in them and each pair of nodes can benconnected by multiple branches with arbitrary time delays. A generalizednlearning algorithm is derived for the ULNs, in which both the firstnorder derivatives (gradients) and the higher order derivatives arenincorporated. One of the distinguished features of the proposed controlnmethod is that the impulse response of the systems is considered as annextended part of the criterion function and it can be calculated bynusing the higher order derivatives of ULNs. By using the impulsenresponse as the criterion function, nonlinear dynamics with not onlynquick response but also quick damping and small steady state error cannbe more easily obtained than the conventional nonlinear control systemsnwith quadratic form criterion functions of state and control variables