Portal de Periódicos da CAPES

Recognition of Radiation Sources for Satellite Communication Based on LSTM

2023; Springer Science+Business Media; Linguagem: Inglês

10.1007/978-981-99-3951-0_89

1876-1119

Kailuo Li, Xiaopo Wu,

Radar Systems and Signal Processing

Recognition of radiation sources remains a higher priority mission for the battlefield intelligence support and even the electronic countermeasure. Recent space networking and the ubiquitous military encrypted satellite communication (satcom) makes the acquisition of measurement and signature intelligence (MASINT) for satcom individuals quite a necessary, which does not depend on the decryption of 0/1 bit stream or priori knowledge of communication messages. This work is dedicated to the recognition of the radiation sources for satcom by virtue of specific phase modulation signals to enhance the MASINT acquisition. A novel classification for satellite communication terminals based on Long Short-Term Memory (LSTM) deep learning method is proposed to explore the underlying features of transmission signals. By utilizing directly the normalized sampling data and designing two hidden layers of multi-memory nodes network, one well-suited deep architecture for nearly real time temporal sequence processing is finally figured out. The impact of arrangement and segmentation of the dataset on the model training is also investigated. Finally, the experimental results have demonstrated the encouraging performance of proposed method.