ASTS: Attention-based spectrum truncation synthesis for step frequency signals

With the development of wireless communication technology, various special pulse waveforms are widely used in radar technology to meet specific requirements. The stepped frequency (SF) signal has properties of low instantaneous band-width, high synthetic bandwidth, and high resolution, which is popular in high-precision target detection. Therefore, researches on SF signals have attracted extensive attention recently. In this paper, we introduced an attention mechanism into spectrum synthesis and finally proposed a novel method called Attention-based Spectrum Truncation Synthesis (ASTS) to solve the problem of energy fluctuation caused by spectrum aliasing and our ASTS method performs better than spectrum superposition synthesis (SSS) algorithm in Peak Sidelobe Ratio (PSLR). Furthermore, we found a noticeable phenomenon in detection caused by a dramatic phase jump happening with the RSF signals when detecting the moving targets. Through the theoretical analysis, we concluded formulas indicating how we could compensate for this phase jump and make the RSF achieve a similar PSLR performance as the LSF does.