Multiple Signal Classification Method of EMF-Based Speed and Position Estimation for Surface PMSM Sensorless Drives

A speed and position estimation method based on Multiple Signal Classification is proposed, which avoid the influence of harmonic components and sliding-mode chattering problem by decomposing the noise and the back electromotive force fundamental component. By utilizing the orthogonality between noise and fundamental wave, the covariance matrix of the signal matrix is subjected to spectral decomposition, resulting in eigenvectors corresponding to noise and fundamental component, respectively. The eigenvectors contain the variation of electrical angle between adjacent sampling points, which can be used to calculate the velocity information with low noise. Then, the electrical angle obtained from the tangent function using the velocity is corrected. Finally, an improved method is proposed to reduce the computational complexity and improve real-time performance.