Intelligent Recognition Method of Radar Pulse Signal Based on Deep SVDD

Le résumé fourni par la source

Aiming at the problems of feature extraction difficulty and low recognition rate caused by the sparsity of radar pulse signal, this paper proposes an intelligent recognition method of radar pulse signal based on deep support vector data description (Deep SVDD). Firstly, the in-phase/quadrature sampling data are preprocessed, and the input real matrix containing both time and frequency domain information is obtained through discrete Fourier transform (DFT) and energy normalization, which significantly simplifies the process of extracting signal features. Then, a deep learning network architecture for radar pulse signal is designed, which includes the residual network, the self-encoder pre-training model and the Deep SVDD network model. Finally, the effectiveness of the proposed method is verified by simulation experiments.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.