| 摘要: | 本計畫以開發自主性導航衛星動態定位演算法為基本訴求；其他比較用的軟體工具為輔，如瑞士Leica廠的LGO程式集。學術研究重點在於：少筆量測資料、多頻電碼與載波相位、接收天線運動中、整數化相位參數及求定位置與速度狀態。因為筆數少，實數相位模稜協方差矩陣橢性極高(長軸和短軸差異很大)，故所提的方差分量解法和解相關變換至為根本。配合著詳盡的公式推導，亦列舉成果圖，為之呼應。就探索新演算知識而言，文獻回顧徹底，形同踏上易行的通道，有利於推動計畫，故列它為常年例行的工作。為方便驗算，實測動態GPS相位資料之建檔與管理工作亦很重要；類比報告繳交，皆屬例行工作。除此之外，依優先順序定義了每年的專業研發項目。第一年期包含動態OTF演算，多頻信號組合與時間序列分析，將實驗性ManGo電腦程式集進一步紮根。第二年籌劃縱橫向工作，即網形大氣改正，多路徑抑制與單點定位；這使得衛定位本領更實在且多元。第三年納進中程、前瞻的狀態參數估計工作，貝氏統計與先驗資訊回饋工作，以期完成定位精度和效度皆屬上乘的總演算法。 ; This project is aimed at developing on our own an algorithm for kinematic navigation satellite positioning, requiring other comparative software tools like the LGO by Leica from Switzerland. Academic research emphasis consists in: a dataset of a few measurements, multi-frequency code and carrier phases, a receiving antenna in motion, integer-valued phase parameters, and the determination of position and velocity states. Real-valued phase-ambiguity covariance matrix is extremely hyper-ellipsoidal because measurements are limited to just a few sampling epochs, meaning that the major axis and minor axis are very different in length. Therefore, the proposed variance-component estimator and decorrelating transformation technique are fundamental in nature. Detailed derivation of equations is accompanied by relevant graphical results. As far as new knowledge on the algorithm is concerned, extensive literature review can be compared to an easy-traveling channel, facilitating the execution of the project. So the review has to be a routine action item. It is also important to set up a data base storing kinematic GPS phase in-situ measurements, for algorithmic validation. Data management and a final report are both annual, routine jobs. In addition, specific research topics are prioritized and listed chronologically. In the first year, an on-the-fly algorithm, the combination of multi-band signals and the analysis of time series are dealt with, solidifying an in-house computer program package, called ManGo. Items intended for the second year are network-based atmospheric correction, multipath mitigation, and precise point positioning. The items make satellite positioning skills become versatile. For the third year, medium-term future-oriented action items involve state parameter estimation, Bayesian statistics, and the positive feedback of prior information. All these efforts can lead to a comprehensive algorithm targeted at top-grade positioning accuracy and efficiency. ; 研究期間 9708 ~ 9807 |
