A New Method For Doppler Sensitive Signal Detection Based on Fractional Correlation

摘要

The Pseudo Random Noise (PRN) signal has a wide application in active sonar detection, because it has good doppler and range resolution. Aimed at traditional matched filter processing of a PRN signal, there is a need for multiple copy signals, because the PRN signal is sensitive to doppler frequency shift. A new detector based on Fractional Correlation is proposed in this paper. As a generalized form of the Fourier Transform Fractional Fourier Trans form (FRFT) can not only deal with the signals of the general sense, but they can also resolve time-varying non-stationary signals. Fractional Correlation processing expands signal processing in Fourier Transform domain to Fractional Fourier Transform domain, hence there is a positive significance in the introduction of FRFT to underwater acoustic detection; and, in the development of a novel detector based on FRFT to deal with complex underwater ambient noise and suppression doppler effects. The theory of a Fractional Correlation detector is derived, and the following are explicitly studied: the detction performance in white Gaussian noise environment; the optimal rotation parameters of the Fractional Correlation detector; and, the computational complexity of the Fractional Correlation detector. The anti-doppler property of Fractional Correlation detector, and the optimal rotation parameter fit to suppress doppler effects are specifically derived. Computer Simulations are provided in order to illustrate the proposed performances of the Fractional Correlation detector. Sea trial data analysis shows that a Fractional Correlation detector base on Fractional Fourier Transform has a better performance than a traditional matched filter, and that it also has good robustness. The Fractional Correlation detector can suppress doppler effects by estimating the optimal rotation parameter in Fractional Fourier Transform domain when the target doppler frequency shift is 2.0 percent of center frequency. The Fractional Correlation detector also has a good performance, so the detector base on Fractional Fourier Transform has broad application prospects.