Sensorless Position Control using Adaptive Wavelet Neural Network for PM BLDCM

In this paper, the relationship between rotor position and phase voltages of permanent magnet (PM) brushless DC motors(BLDCM) is analyzed, a novel sensorless position control method for BLDCM is proposed, which is based on adaptive wavelet neural network(WNN). In this method, a multiple inputs-single output wavelet neural network model is built according to the analysis of sensorless position and micro-controller. There is no hidden unit at the beginning. During the learning process, the hidden units are increased or decreased according to an adaptive algorithm. The WNN is finally built with a much simpler and more compact structure. Through a gradient descent error algorithm, the WNN is trained both off-line and on-line. In the off-line training, the parameters of network are initialized and the number of hidden units is obtained. While on-line training, the weights of WNN are updated based on the output signals after filter. By mapping the phase voltages to commutating signals, the network can replace the traditional position sensors. By the experimental results, the method in this paper is verified to get the exact commuting signals.