Correlation Structures and Prediction Models for Radar Signal Values of Sea Waves

Abstract

A naval radar, though designed to detect objects or targets, also receives backscatter from wave peaks. Better understanding of the signals generated by the background waves can improve the detection of small targets. This thesis examines the correlation structure of the sea waves along the dimensions Doppler, range, azimuth and scan. The targets are mathematically simulated, as are the sample waves used alongside the measured waves. Using two prediction models – Linear Prediction and Sequential Decorrelation – examination of each cell illustrates the difference between predicted and measured signals in a plot of residual signals. High values in the resulting plot indicate target presence. Further suppressing the wave signals enhances anomalies caused by target signals via discrimination and calibration techniques. The findings matter to those wanting clearer target detection on a wave-covered sea surface situation.