Simulation of Correlated Low-Grazing-Angle Sea Clutter Based on Phase Retrieval

It is known that the major difficulty of sea clutter simulation is the controlled generation of a continuous correlated non-Gaussian random process. In particular, for sea clutter at low grazing angles and over a long time scale, the spiky and nonstationary nature makes the simulation more difficult. This paper proposes a novel procedure for the simulation of continuous correlated low-grazing-angle sea clutter with prescribed statistic and correlation characteristics. The Pareto distribution is utilized to describe the statistics of the sea clutter intensity. In addition, different correlation characteristics for sea clutter over both short and long time scales are considered in this paper. First, the memoryless nonlinear transform is adopted to simulate the intensities of the correlated sea clutter. Second, the Doppler spectra of different range bins are generated separately. In particular, for the sea clutter over long time scales, successive time-varying Doppler spectra are modeled to characterize the nonstationarity of the sea clutter series. Then, the phases of the sea clutter are retrieved by constraining to the desired Doppler spectra and given magnitudes. Different methods are adopted for the phase retrieval of sea clutter on short and long time scales. Finally, simulation results show that the proposed procedure can generate continuous low-grazing-angle sea clutter with prescribed statistic and correlation characteristics. In particular, for the sea clutter over a long time scale, the simulated sea clutter series enjoys the time-varying Doppler spectrum while maintaining continuity. In addition, the proposed method is also suitable for other models, such as the K-distribution, the Weibull model, and so on.