Abstract
Due to V-belt continuously variable transmission (CVT) driven electric scooter with unknown nonlinear and time-vary characteristics, an accurately dynamic modeling of V-belt CVT driven electric scooter is the time-consuming procedure for the linear controller design. A hybrid modified recurrent Legendre neural network (NN) control system is proposed to control the V-belt CVT driven electric scooter under the occurrence of nonlinear load disturbances with parametric variations in order to acquire good control performance. Firstly, system structure of the V-belt CVT driven electric scooter by using a permanent magnet synchronous motor (PMSM) servo drive system is proposed in this study. Secondly, the hybrid modified recurrent Legendre NN control system, which consists of an inspector control, a modified recurrent Legendre NN control with adaptive law and a recouped control with estimated law, is proposed to control the V-belt CVT driven electric scooter by using PMSM servo drive. Moreover, the on-line parameters tuning methodology of the modified recurrent Legendre NN is derived according to Lyapunov stability theorem and gradient descent method. Furthermore, two optimal learning rates of the parameters in the modified recurrent Legendre NN are developed to speedup convergence. Finally, to show the effectiveness of the proposed control scheme, comparative studies are demonstrated by simulated and experimental results.