利用數值地表模型進行衛星定位品質評估

Abstract

衛星定位相關技術已發展數十年，最早是以軍事用途為主要目的，用以提供精確的定位及定時能力，而除了用於精密的定位測量外，用於交通工具上之導航及其他民生工業之產品也相當普及。此外，不同的衛星定位系統亦相繼發展中，未來地面之覆蓋率將大幅增加，並使這項技術之可靠度及應用性更進一步提昇。 衛星定位測量目的在於求得地面點位之精確的三維位置，而衛星的可視性以及空間分佈是影響定位品質的關鍵因素，因此測量前的點位預先規劃與衛星可視條件分析變得相當重要。現有測量規劃軟體之中多數僅模擬平面地形或以設定遮罩角之方式進行定位品質預測，並無法真實反應現地的地形遮蔽效應。考量近年來空間資訊技術不斷快速發展，高品質的地形資訊變得愈來愈容易取得，因此本研究將利用高精度之數值地形模型來評估模擬地形效應及真實之衛星定位品質。 使用高解析度地形資訊時，為增加地形資料之分析效率，本研究提出一項自適應取樣之演算方法，透過合理的取樣間隔，在滿足預先設定的分析精度條件下，使分析速度大幅增進。另外，由於衛星軌道必然包含不確定性，而傳統方法中將衛星位置視為常數，並沒考慮其精度，而本研究則採用非固定約制平差模型，推導考量衛星坐標之精度之定位品質計算方法，以使得精度評估成果更為合理可靠。 為驗證本研究所提出方法之可靠度及效能，本研究使用高精度數值地表模型進行衛星測量規劃與實測作業，在實驗區中比較模擬分析與真實測得之衛星可視性，實驗成果顯示本研究之方法能正確地預估現地的施測時衛星可視條件，對於提升衛星定位品質評估的可靠度將有具體助益。

In these decades, satellite positioning is becoming a very important technique applied in our daily life. The basic principal of this technique is to determine a receiver’s location using range observations between satellites and the receiver. Therefore, the intervisibility between satellites and receivers is one of the most critical factors to the positioning quality. Typically, satellite positioning with a low obstruction will be more accurate and is thus preferred. The positioning quality can be pre-analyzed by identifying all visible satellites to a specific receiver. However, the topographic consideration in conventional planning software is usually neglected or simplified as a mask angle. Consequently, the results predicted by conventional software will not be realistic since detailed topographic data is not involved. Fostered by the rapid development of spatial information technique, three-dimensional topographic information (e.g. Digital Terrain Model and Digital Surface Model) of a higher quality is becoming accessible to the public. This study aims to develop a satellite visibility and quality assessment technique utilizing high resolution 3-D topographic information. An adaptive sampling and analysis technique is proposed to increase the computational efficiency while processing topographic data of various grid resolutions. Additionally, a unified least squares approach is introduced to model the uncertainties of satellite orbits. Simulation tests and a GPS field work have been conducted to demonstrate the performance and capability of this approach. The results reveal a significant improvement on the reliability for the quality estimation of a satellite surveying. Consequently, the proposed approach will benefit the applications in which a pre-analysis of the positioning quality is of a major concern in a satellite surveying (e.g. a GPS filed planning or network design).