GNSS連續觀測站遠距頻率校正及定位精度分析

為維護基本控制點品質，確保土地測量精度，本研究以GPS共視法（GPS common-view）開發GPS遠距頻率校正技術。利用相位法與電碼法解算13處GNSS觀測站之時間偏差量（phase offset），並以頻率穩定度（frequency stability）與頻率偏移（frequency offset）進行精度驗證，所得到的精度有11站與GPS衛星規格（10-12~10-11）相同，顯示GPS遠距頻率校正技術已可應用於台灣測站，未來將透過遊校（on-site calibration）直接比對進行驗證。惟CIME與WARO兩站之頻率穩定度與衛星規格相差3~5個等級，推斷為其內部石英震盪器異常所致。頻率穩定度及頻率偏移對精密單點定位（Precise Point Positioning, PPP）精度的影響大於對靜態相對定位（Static Relative Positioning），由於靜態相對差分定位消除時錶誤差，使頻率穩定度最差的CIME（10-7）及次差的WARO（10-8）其靜態相對定位精度在平面及高程平均達2.4 mm與6.6 mm；而採精密單點定位時，CIME定位精度在平面及高程僅10.1 mm與29.9 mm，顯示當接收儀頻率誤差大時，採用靜態相對定位對維持土地測量的精度較有利。從解算時間對定位精度的影響來看，靜態相對定位在解算時間2小時以上精度即顯著提升（~66%），24小時平面及高程精度最佳達1.1 mm與2.9 mm；而精密單點定位需大於4小時的解算時間方能提升精度，其24小時平面及高程精度最佳為1.9 mm與2.9 mm。

This study aims to maintain the quality of basic control points and ensure land survey precision by devising GPS remote frequency calibration techniques. This study calculates time offset for 13 GNSS observation stations using both the phase method and the code method. The precision is validated through assessments of frequency stability and frequency offset. The resulting precision aligns with that of GPS satellite specifications (10-12~10-11), affirming the applicability of GPS remote frequency calibration techniques to stations in Taiwan. In the future, it will be verified directly through on-site calibration. Notably, the frequency stability of the GNSS observation stations located at CIME and WARO exhibit considerable discrepancies, which is 3 to 5 orders higher than satellite specifications, thereby implying abnormalities within their internal quartz oscillators. The impact of frequency stability and frequency offset on precise point positioning (PPP) precision is greater than that on static relative positioning precision. Since static relative positioning eliminates the clock error, the positioning precision of WARO and CIME with the worst frequency stability (10-8~10-7) is still good (horizontal 2.4 mm and vertical 6.6 mm). It shows that static relative positioning is more beneficial to maintain the precision of land survey than PPP when the frequency error is large. The precision of static relative positioning is significantly improved (~66%) when the time period of data is 2 hours. Computing of 24 hours data can reach the positioning precision to 1.1 mm and 2.9 mm in horizontal and vertical, respectively. However, the precision of PPP is improved when the time period of data is 4 hours. The positioning precision in horizontal and vertical are up to 1.9 mm and 2.9 mm, respectively, when computing by 24 hours data.