Resolver Decoding Method Using Hilbert Transform and ATO

Abstract

Resolvers are widely used in electric vehicles, trains, and other harsh fields because of their robustness. However, the resolver outputs two orthogonal analog signals, which make the resolver decoding either high hardware cost or poor decoding accuracy. A noise robust resolver decoding method using Hilbert transform and angle-tracking observer (ATO) is proposed in this paper. Firstly, Hilbert transform is employed to obtain the modular envelopes of resolver signals. Next, the modular envelopes are filtered, and their quadrants are recognized by the polarity relation of the resolver signals and the modular envelope extreme point. Then, the ideal demodulating signals are gained through the linearization of the envelope zero point. Finally, the improved ATO is used to obtain the rotor angle by iteratively calculating the demodulating signal. The effectiveness of the proposed method is verified by experiments under various rotor speed conditions and compared with other methods in noise immunity. The results show that the proposed method can control the decoding error within 0.5° when the SNR is 30 dB, which provides a high-precision and low-cost decoding scheme for practical applications.