Mitigating Non-Target Reflections in GNSS-IR Using a Low-Cost Directional Antenna
IEEE Access, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1109/access.2026.3742223
- Dergi Adı: IEEE Access
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Directory of Open Access Journals
- Anahtar Kelimeler: directional antenna, GNSS-IR, reflectometry, sea level, signal-to-noise ratio
- Karadeniz Teknik Üniversitesi Adresli: Evet
Özet
Global Navigation Satellite Systems Interferometric Reflectometry (GNSS-IR) retrieves environmental variables from the interference between direct and surface reflected signals. Unwanted contributions are commonly removed by azimuth masking, which acts on the computed satellite azimuth rather than on the direction of arrival, so reflections from outside the selected sector remain in the recorded signal-to-noise ratio data and cannot be separated afterwards. Data from HNLC, a coastal GNSS station next to a strongly reflecting roof, illustrate this limitation. The novelty is hardware restriction of the angular response, at the aperture and before recording. A low-cost 3D-printed Dual Ridged Horn Antenna (DRHA) concentrates sensitivity in the sector of the target surface and suppresses arrivals from behind. It was deployed at a fixed reflecting surface with a low-cost receiver and compared against a commercial patch antenna in its conventional zenith-up orientation on the same receiver, under the same azimuth mask, with single-frequency observations from three constellations. The DRHA returned a median of 2.991 m against an in-situ reflector height of 2.99 m, with a median absolute deviation of 0.226 m consistent across GPS, GLONASS and BeiDou, while the patch median was displaced by 1.538 m from its own in-situ height, with a five times larger deviation. Median peak-to-background noise ratios were 4.89 against 3.73 overall and 4.59 against 3.78 over 16 matched GPS pairs (Wilcoxon signed-rank, p = 0.056). As a proof of concept, hardware-level angular selectivity is a practical, low-cost complement to masking where the target cannot be isolated in the data domain.