Research On The Doppler Velocity Measurement Post Processing Technology Based On Maximum A Posteriori Estimation

The technology of Doppler velocity measurement is an important high-precision active speed measurement technology,widely used in marine resources development,ocean observation,waterway measurement and many other fields.The accuracy of velocity measurement is an important index of Doppler velocimetry technology.In this paper,we aim to improve the velocity accuracy and mainly study the application of maximum a posteriori estimation in Doppler velocimetry.This paper firstly combines the coherent velocity measurement technique with the Maximum a posterior estimation to complete the correlation analysis of the algorithm.For measurement models,echo signal models based on point-scattering signal superposition are studied.Due to the large amount of computation,echo signal models based on Gaussian power spectral density are studied.Based on this,the correction of correlation coefficient and phase distribution are studied.For the correction of the correlation coefficient,the asymptotic formula of the infinite pulse pair is deduced.When the finite pulse pair is used,the Gaussian echo signal model is used to modify the correlation coefficient and the validity is verified.For the phase distribution,the formula of the phase distribution under a single pulse pair is deduced.The Gaussian echo signal model is used to study the phase distribution under multiple pulse pairs,and the phase distribution of the Gaussian echo signal model is verified by the point-scattered echo signal model.The first-order AR model was used as a prior model to determine the prior probability density function.On the basis of forward filtering and backward filtering,smoothing is used to eliminate the estimation delay,and a method for determining the smoothing parameter is given from the perspective of power spectrum.The data processing results show that the power spectrum of the processed velocity sequence satisfies the-5/3 slope within a certain interval,and the standard deviation is also greatly reduced,which shows that the algorithm can effectively reduce the noise and improve the speed accuracy,and achieve the expected goal.