The single neuron PID controller for automotive electromagnetic clutch based on RBF NN recognition

It is more and more popular to use electromagnetism clutch on mini-cars, for its unique structure and simple controlling. But the drive of the electromagnetism clutch has some nonlinearity, and there is interference around it, so the traditional PID control strategy couldn´t work at its best. The single neuron self-adaptive PID controller using on-line identification of radial basis function Neural Networks is proposed in this paper, which is used to control the current of the clutch. The single neuron is applied to the self-adaptive controller for electromagnetic clutch, for it has such abilities as self-study and self-adaptive. Meanwhile through on-line recognition of the system by RBF NN, the parameters of the controller are set on-line. The simulation result proves that the dynamic performance of electromagnetic clutch driver is better, and the system has nice adaptability.