Classification of radar clutters with Artificial Neural Network

The detection performance of radars is usually limited by clutters. Clutter models are necessary factors in the simulation and modeling of radar systems. Therefore, Knowledge of clutters is important to identify targets and The impact of clutter cannot be ignored. Statistical distributions are established models which are widely used in radar clutters performance calculations. Gaussian, Weibull, Rayleigh and K-distribution are four common clutter models. The design of radar detectors and clutter classification are really complicated issues. For clutter classification, the different methods such as Bayesian classifier are used. But in this paper we propose a new classifier for radar clutters. We use Artificial Neural Network to classify four common radar clutter models. First we simulate clutter models then extract their important features like skewness and kurtosis. Finally, we make neural network with two hidden layers which has twenty and fifteen neurons in each layer and we separate clutter models with high accuracy. The results prove the validity of proposed method.