Linearization control of VCR servo system by using a neural network based feedforward compensator

Using the fact that a neural network can emulate a nonlinear function, we design a simulated neural network compensator for VCR servo system. The designed model is analyzed theoretically and the performance of the model is tested through extensive simulations by applying the model to a real commercial product (VCR SV-H815). Simulation results show that our controller reduces steady state velocity error by 60% and low frequency wow-flutter problem is reduced considerably.