Abstract: For the stability problem of spatial spectrum estimation based on the minimum variance distortionless response (MVDR) method, a kind of spatial spectrum estimation method based on coherent cumulative preprocessing is proposed. First, complex analytic data with a certain frequency band is transformed by complex analytic wavelet transform from the collected data of receiving array, and the sub-group data is obtained according to the idea of spatial smoothing from the complex analytic data. Next, making full use of the phase information of sensor data, a group of data with high signal-to-noise ratio is obtained by accumulating the complex analytic data of each sub-group after time delay compensation. Then, the covariance matrix of the new data is constructed via multi-point cumulative processing in time domain. Finally, spatial spectrum estimation is realized according to the orthogonal property of the covariance matrix. The processing results of numerical simulation and measured data show that, compared with the MVDR method and the diagonal loading MVDR method, the data source of constructing the covariance matrix is changed in this method through time domain complex analytic transform and coherent accumulation pretreatment. The full rank covariance matrix is stably obtained by multiple sampling points accumulation in this method. According to the relation of spatial bearing and the phase difference of the sensor data, this method can effectively improve the stability of spatial spectrum estimation via double exponential function addition.

Key words: spatial spectrum estimation; complex analytic transformation; coherent accumulation; multipoint accumulation