Task Sequence Filling for Multi-function Radar via BERT Model

The received multi-function radar (MFR) pulse signals usually are partially lost in the complex electromagnetic environment (e.g., multiple radiation sources scenario, low signal noise ratio (SNR) scenario), making it difficult for the reconnaissance team to correctly sense the MFR task sequence which further affects the accuracy of electronic intelligence information. To this end, this paper develops a MFR task filling method based on the BERT model in the presence of complex electromagnetic environment. More precisely, a MFR hierarchical model is established which consists of waveform unit, radar task and work mode. Exploiting the radar knowledge database, the intercepted pulse sequence is mapped to the radar task sequence, and the BERT model is further introduced to fill the radar task sequence. The simulation results show that the proposed methodology achieves relatively high filling accuracy in different working modes.