Abstract:

Locating mobile devices in real time is crucial in electronic countermeasures systems, whose performance mainly depends on the estimated speed of the direction of arrival (DOA). Low snapshots are a prerequisite for fast DOA estimation. The current DOA estimation based on the sparse reconstruction algorithm has the advantage of adapting to low snapshots, but the estimation accuracy is limited by the initial observation matrix, and the estimation speed is limited by multiple iterations of the high-dimensional observation matrix. To this end, a fast estimation algorithm with high precision under low snapshots coupled with spatial difference matrix and sparse reconstruction is proposed. Firstly, the spatial difference matrix is used to eliminate the influence of incoherent signals and noise on the estimation results of coherent signals, and improve the accuracy of the initial observation matrix. Then, the complete dictionary is processed with spatial smoothing before and after, so as to overcome the complex problem of high-dimensional signal processing and realize fast estimation. Finally, the incoherent and coherent signals are estimated separately. The simulation verification result shows that compared with the sparse reconstruction method, the sensitivity of the initial value of the proposed scheme is significantly reduced, and the running time complexity is reduced by more than half under the premise that the accuracy is guaranteed to be equal or even slightly improved.

Key words: direction of arrival (DOA) estimation, sparse reconstruction, spatial difference, mixed signal, spatial smoothing