Wind Profile Estimation for K-Band Doppler Radar Based on Mutual Convolution Cost Function Method

At present, K-band Doppler radar has been proposed for detecting the motion state of atmospheric turbulence and obtaining wind velocity and direction. However, the echo signal of the turbulence is often interfered with strong noise, and there are multiple spectral peaks in the power spectrum, which greatly affect the overall detection accuracy of the radar. Traditional wind profile estimation methods directly analyze the coherent accumulation (CA) signal using a cost function, which is susceptible to clutter and noise during the selection of the candidate spectral peaks. To overcome the shortcomings of traditional methods under the condition of low signal-to-noise ratio (SNR) and multiple peaks, in this article, we propose a novel method based on the mutual convolution cost function (MCCF), which utilizes the mutual convolution and the complex least mean square (MCLMS) method to filter the CA signal, and then generates a cost function to select the optimal spectral peaks in the space-time Doppler window. This systematically solves the problem of turbulence echoes being interfered with strong noise and multiple peaks. Finally, the commonly used methods are compared with the proposed method, and the statistical results of these methods are verified by using LIDAR data. The results indicate that the proposed method can improve the SNR by 2–4 dB and improve the overall correlation coefficient (CC) by about 10% of the radar.