Adaptive signal accumulation processing systems have found significant application due to a priori uncertainty in signal separation against the background of passive noise. At the same time, when synthesizing Doppler phase meters, it is appropriate to adapt not only to noise parameters, but also to the Doppler phase of the signal. Adaptation to the signal in such systems is not carried out in the event of a significant excess of the noise level by the signal, which occurs in many cases. In the study, it was required to obtain a quasi-optimal algorithm for estimating the Doppler phase of a signal and analyze it, as well as to study the properties of the obtained estimates from the point of view of their use in adaptive signal accumulation. In this case, the procedure for overcoming the a priori uncertainty of noise parameters used an adaptive Bayesian approach and the obtained consistent noise estimates. In turn, the calculation of these estimates is based on the use of a recurrent procedure and algorithms for maximum likelihood estimates. An optimum linear filtering algorithm was obtained for the signal additively mixed with passive correlated noise. Due to the uncertainty of the Doppler phase shift of the signal, it is shown that the processing system must have N channels. In turn, such a solution makes it possible to limit the uncertainty of the Doppler phase shift of the signal to the bandwidth of each channel. The algorithm itself is implemented as a non-adaptive multi-channel filter. Using the output samples of the adaptive matrix filter used serves to overcome the difficulties of adapting to the signal parameters. Matrix processing also provides sufficient noise cancellation and adaptation to the Doppler phase of the signal. A quasi-optimal algorithm for estimating the Doppler phase of the signal is also obtained, in which averaging is carried out based on readings of one resolution element. The structural diagram of the meter corresponding to this algorithm is given. Measurement accuracy of quasi-optimal algorithm is determined and shift of Doppler phase estimate of signal is found. The bias is related to the frequency response shape of the adaptive matrix filter. There is no offset in the filter bandwidth, and outside it is determined by the shape of the noise spectrum and the noise/interference ratio. The results of simulation of mathematical expectation and standard deviation of the said estimate under the influence of noise with the Gaussian correlation function are given. The possibility of using the obtained estimates in processing systems with adaptive signal accumulation, having greater accuracy of estimating the Doppler phase of the signal, is shown.
Artyushenko et al. (Wed,) studied this question.