Model based coherent detection in medium grazing angle sea-clutter

The time-varying and range-varying nature of the sea clutter Doppler spectra presents a challenge to coherent detection performance. Detectors using a large number of surrounding cells to estimate spectra or covariance matrices often give poor control of false alarms. Conversely, attempts to use a smaller number of possibly more representative range cells can result in numerical stability problems. In this paper, we introduce a model based spectral estimate and show how the false alarm rate can still be well controlled, while using data from only a small number of surrounding range cells. A number of implementations are presented for both the time (pre-Doppler) and frequency (post-Doppler) domains. These methods require only a small number of range cells to work and have a reduced computational load, making them suitable for implementation in real radar systems.