Sea Clutter Sequences Regression Prediction Based on PSO-GRNN Method

In marine radar signal processing, in order to suppress sea clutter, sea clutter sequences regression prediction is necessary. Sea clutter has chaotic features, and GRNN (General Regression Neural Network) algorithm can effectively predict regression of chaotic sequences, this paper presents a sea clutter sequences regression prediction method based on an improved GRNN algorithm, using phase space reconstruction to strike GRNN training samples, applying adaptive PSO (Particle Swarm Optimization) algorithm to optimize GRNN Gaussian width coefficient, then the IPIX radar data of Canada Mc Master University were used, to doing the experiment on sea clutter forecast. The results showed that: regression model to predict sea clutter is feasible, and PSO-GRNN method can higher improve the prediction accuracy than GRNN method.