Underdetermined Source Signal Reconstruction Based on Multi-modal Fusion

In complex electromagnetic environments, non-cooperative wide-band receivers encounter the challenge of separating and identifying unknown source signals from time-frequency aliased signals. When the number of received channels is smaller than the number of source signals, it leads to the under-determined signal separation problem, making the task particularly challenging.To circumvent this limitation, this paper proposes an under-determined source signal reconstruction method based on multi-modal fusion (URBM).The proposed method effectively utilizes the complementary characteristics of different modal signals such as time-frequency maps and wave-frequency maps, thereby improving the accuracy and robustness of source signal reconstruction.