YONLAMA SONAR MA’LUMOTLARI BO‘YICHA SUV TUBIDAGI OBYEKT KOORDINATALARINI ANIQLASHNING SONLI USULLARI
Keywords:
Kalit so‘zlar: yonlama sonar, moslashtirilgan filtr, lokalizatsiya, Gauss–Nyuton usuli, Monte-Karlo, Kramer–Rao chegarasi., Keywords: side-scan sonar, matched filter, localization, Gauss–Newton method, Monte Carlo, Cramér–Rao bound.Abstract
Annotatsiya. Maqolada yonlama sonar ma’lumotlari bo‘yicha suv tubidagi nuqtaviy obyekt koordinatalarini aniqlashning sonli usullari ishlab chiqilgan va taqqoslangan. Qiya masofa chiziqli chastota modulyatsiyali (LFM) signalni moslashtirilgan filtrlash va korrelyatsiya cho‘qqisini parabolik interpolyatsiya qilish orqali, bo‘ylama koordinata esa aks-sado amplitudasining impulslar bo‘yicha profili asosida uch usul — maksimal impuls, markaz (centroid) va Gauss–Nyuton usulidagi eng kichik kvadratlar — bilan baholandi. Monte-Karlo tajribasida (kirish SNR −24 … 0 dB) −12 dB va undan yuqorida aniqlash ehtimoli 1,00 ga teng bo‘lib, qiya masofa xatoligi Kramer–Rao chegarasidan ±9 % dan ko‘p chetlashmadi (−16 dB dan yuqorida). 0 dB da bo‘ylama koordinata RMSE: 6,7 sm (maksimal impuls), 1,0 sm (markaz), 0,9 sm (Gauss–Nyuton).
Abstract. Numerical methods for determining the coordinates of an underwater object from side-scan sonar data. Numerical methods for estimating the coordinates of a point object on the seabed from side-scan sonar data are developed and compared. Slant range is estimated by matched filtering of a linear-FM signal with parabolic peak interpolation; the along-track coordinate is estimated from the ping-to-ping echo amplitude profile by three methods: peak ping, centroid and Gauss–Newton least squares. In a Monte Carlo study (input SNR −24 to 0 dB) the detection probability is 1.00 from −12 dB upward and the range error stays within ±9 % of the Cramér–Rao bound above −16 dB. At 0 dB the along-track RMSE is 6.7 cm (peak ping), 1.0 cm (centroid) and 0.9 cm (Gauss–Newton).
