Radar detection algorithm for GARCH clutter model

We propose a GARCH model to represent the clutter in radar applications. We fit this model to real sea clutter data and we show that it represents adequately the statistics of the data. Then, we develop a detection test based on this model. Using synthetic and real radar data, we evaluate its performance and we show that the proposed detector offers higher probability of detection for a specified value of probability of false alarm than tests based on Gaussian and Weibull models, especially for low signal to clutter ratios.