IDENTIFICATION OF BELT-DRIVEN SERVOMECHANISM USING CHEBYSCHEV FUNCTIONAL RECURRENT NEURAL NETWORK

摘要

This paper proposes Chebyshev functional recurrent neural network (CFRNN) as a novel functional link neural network to identify the nonlinear system. The convergent analysis of CFRNN is discussed in detail. The nonlinearity is introduced into CFRNN by enhancing the input pattern with nonlinear functional expansion of Chebyshev polynomial. A flat belt servomechanism is taken into consideration, which is identified by CFRNN and the result is compared with that of recurrent neural network.