Online unsupervised generative learning framework based radar jamming waveform design

The jamming effect on radar is dominated by the design of waveform directly. Traditional waveforms are generated using template-based method and not environmentally resilient. Without precise centre frequency support, direct guidance signal power spectrum density (PSD), a novel with an online unsupervised generative learning framework proposed. proposed consists two modules: generation module optimisation module. For module, well-designed neural network (NN) used to generate waveform. While for iterative loss functions developed as tasks guide back-propagation online. Given transmit power, PSD designed adaptively focus band fit well, which improves receiver in-band jamming-to-signal ratio (JSR). It worth noting that NN end-to-end trained without extensive prior training data. can within reasonable physical parameter ranges, numerical experiments have shown performed better effects target detection in comparison traditional one.