A novel time-frequency atomic dictionary for radar intra-pulse modulation signal sparse representation

The sparse representation using time-frequency atoms is a useful analysis method for radar signal processing. Based on Time-Frequency theory, time-frequency atoms can match radar signals and represent it sparsely. Traditional atomic dictionary can only match the linear frequency modulation signal, while a novel time-frequency atomic dictionary proposed in this paper can not only match the linear time-frequency modulation signal properly but also match the non-linear time-frequency relationship appropriately. It is because that the proposed atomic dictionary has a sinusoidal factor which improves the atoms' ability to match and represent the linear and non-linear frequency modulation signals. Compared with traditional atomic dictionaries, the simulation result shows that the novel atomic dictionary can represent radar intra-pulse modulation signals more effectively and sparsely.