Direction Finding of Correlated Sources in Extremely Low SNR Environment | ||||
| Journal of Engineering Science and Military Technologies | ||||
| Articles in Press, Accepted Manuscript, Available Online from 25 February 2024 | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/ejmtc.2024.218487.1259 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
Osama Ahmed Elkasrawi  1; Amr Mohamed Abdelaziz2; Hossam El Din Abou Bakr1; Mohamed Atta Aboelazm1
| ||||
| 1Department of Electronic Warfare, Military Technical College, Cairo, Egypt | ||||
| 2Department of Communication, Military Technical College, Cairo, Egypt | ||||
| Abstract | ||||
| Direction-Finding (DF) in low Signal-to-Noise Ratio (SNR) environments presents challenges in accurate Angle-of-Arrival (AOA) estimation, resolution, and maximum detection range due to extrinsic sources and hardware imperfections, particularly when dealing with correlated sources. Spatial-Smoothing (SS) techniques are developed to solve the problem of fully correlated sources at the expense of potential loss of both antenna aperture and angular resolution. In this paper, we introduce a pre-processing ”oversampling and averaging” stage that works collaboratively with SS stage to overcome the degraded resolution performance when dealing with correlated sources. We show that oversampling intercepted signal, with an over-sampling ratio (OSR), and decimating it back to its original Nyquist rate can significantly enhance the SS-based conventional Multiple Signal Classification (MUSIC) AOA estimator’s ability to resolve fully correlated sources at low SNR and narrow angular separation, which carried over a low-cost hardware environment. We derive the Cramér–Rao Bound (CRB) for AOA estimation, which shows a reduction by the same OSR factor. Simulation results at SNR= −15 dB with OSR=50, ensure a drastic drop in the Root-Mean-Square-Error (RMSE) of the MUSIC-based AOA estimator from 31◦ to 0.5◦. The resolution enhancement is confirmed by sharper peaks in the MUSIC spectrum. Simulation analysis further validates the improvement in the SS method. To demonstrate practical feasibility, we implement a prototype test-bed using the National-Instrument-Universal-Software-Radio-Peripheral (NI-USRP) Software-Defined-Radio (SDR) platform. Experimental results confirm the effectiveness of our proposed approach. Detailed mathematical analyses in the appendices support the derived results and defend our findings. | ||||
| Keywords | ||||
| MUSIC; USRP; RMSE; CRB; Oversampling | ||||
|
Statistics Article View: 39 |
||||
