Prediction of Multi-Function Radar Signal Sequence Using Encoder-Decoder Structure

The multi-function radar (MFR) can perform a variety of tasks and has complicated and adaptable waveform characteristics. Predicting MFR transmission signals is critical for electronic warfare (EW) devices to develop effective countermeasures. This paper therefore deals with the signal sequence prediction problem in the presence of multiple targets via using an encoder-decoder structure. Specifically, the gated recurrent units (GRU) network is developed as the encoder to extract the temporal features, while the transformer decoder layers are exploited as the decoder to output the prediction results. The experimental results demonstrate that the method proposed can achieve effective prediction even in non-ideal situations.