Compound-Gaussian Spatial-Temporal Correlated Complex Clutter Simulation Based on a Two-Step Data-Driven Method

In this work, aiming to simulate random complex clutter with the similar spatial-temporal correlation and amplitude distribution to the measured clutter, a two-step data-driven method as an alternative to statistical methods is proposed. First, a complex data matrix with the desired spatial-temporal correlation is generated via the sliding-window filtering (SWF) from the measured clutter. Based on the compound-Gaussian model, it is explained that the SWF guarantees that the complex data matrix can roughly reproduce the amplitude distribution of the measured clutter. Second, to further improve the amplitude distribution reproducing performance, a scheme combining the amplitude-phase separation technique and the correlation transfer method is introduced to process the complex data matrix generated by the SWF. The experimental results show that the clutter generated by the two-step method can reproduce the main statistical characteristics observed in the measured clutter.