Multi-target Separation Algorithm Based on DFT Transform in Uniform Linear Array Scene
Kai Wang,Jincheng Li,Zichuan Yu,Jianxun Li,Lu Tang,Xusheng Tang
DOI: https://doi.org/10.1109/lsens.2024.3417381
2024-01-01
IEEE Sensors Letters
Abstract:A discrete Fourier transform (DFT)-based multitarget separation method is proposed in this letter. This method can accurately estimate the direction-of-arrivals (DOAs), amplitudes, and phases of far-field narrowband source targets. First of all, the method estimates the signal's DOA by DFT analysis of snapshot data. Second, the method estimates the amplitude and phase of each signal with multisnapshot data. Finally, this approach enables the reconstruction of the time-domain sampling sequence of received signals and separates multisignal. Compared to traditional root-multiple signal classification (Root-MUSIC), estimation of signal parameters using rotational invariance techniques (ESPRIT), non-iterative fast method (NIFM), fast iterative interpolated beamforming (FIIB) methods, and temporal cross-correlation methods, the method proposed in this article demonstrates superior performance in both single and multisnapshot scenarios. Experimental results indicate significant advantages of this method in DOA estimation, signal separation, and reconstruction of received signals, especially suitable for applications in noisy environments. The number of signals supported by this method is closely related to the number of elements in the sensor array, offering good flexibility and scalability according to practical application requirements. By comparing the performance of different methods, this study validates the superiority of this method under various SNR and signal interval conditions, providing valuable insights for future DOA estimation and signal processing research.