Radar signal recognition based on ambiguity function features and cloud model similarity

Abstract

With the purpose to improve the recognition performance of radar emitter signals in case of low Signal-to-Noise Ratio (SNR), the paper presents a method based on main ridge slice of ambiguity function (AF) and cloud model similarity . Based upon the fact that different signals result in different AF structures, the proposed method selects the rotation angle and cloud model similarity coefficients of AF main ridge slice as feature parameters, and constructs the feature vectors for the classification and recognition of radar emitter signals. In the process of feature extraction, Empirical Mode Decomposition (EMD) is employed to weaken the influence of noise on main ridge slice envelope. In simulations, Kernel Fuzzy C-means (KFC) clustering is adopted to realize the recognition of different types of radar signals. Experimental results show that aggregation within class (AWC) and separability between classes (SBC) of extracted feature vector remain good in spite of dynamic SNR. The proposed method is capable of classifying and identifying radar emitter signals with higher correct recognition rate (CRR) in comparison with existing methods.