請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/20541
標題: | 適應性卡爾曼濾波器在單頻GNSS精密單點定位之應用 The Application of Adaptive Kalman Filter In GNSS Single Frequency Precise Point Positioning |
作者: | Yen-Lin Chen 陳彥霖 |
指導教授: | 王立昇(Li-Sheng Wang) |
關鍵字: | 單頻,精密單點定位,卡爾曼濾波器,適應性卡爾曼濾波器,全球衛星定位系統,精進卡爾曼濾波器, single frequency,precise point positioning,Kalman Filter,Adaptive Kalman Filter,GNSS,refinement, |
出版年 : | 2017 |
學位: | 碩士 |
摘要: | 隨著智慧生活與物聯網的崛起,未來對低功率、低耗能衛星定位系統的需求將會更高,因此本研究使用單頻的精密單點定位系統以及全球定位系統(Global Positioning System, GPS),並希望能達到精準即時動態定位之目的。
為了使估測最佳化,觀測量誤差部分使用IGS提供的Ultra-Rapid精密星曆修正GPS的衛星的鐘差與衛星軌道的計算;電離層誤差使用IGS所提供的全球電離層地圖而對流層誤差則是使用UNB3m的模型,兩者皆透過內插的方式來進行修正。 而為了能夠在動態上有更穩定、可靠度更高的定位結果。本研究使用載波平滑偽距(Carrier smooth code pseudorange)與都普勒頻移(Doppler shift)兩個觀測量來進行位置與速度的解算,並且使用適應性卡爾曼濾波器的演算法。在適應性部分採用累積狀態的殘差向量針對系統雜訊共變異矩陣進行估測,而觀測雜訊共變異矩陣的估測則是透過大數據處理的方式進行統計,找出趨勢並且進行迴歸。此外,本研究嘗試進行卡爾曼濾波器的精進,將估測結果重新進行微調。經過比較後可以發現不管在動態或靜態實驗中,適應性的效果對於定位結果有明顯的提升,而精進部分則對動態實驗中有不錯的效果。 With the booming development of Internet of Things (IoT), the market demand for low-cost and low-energy-consuming positioning system will be grown up in the future. Therefore, in this work, we try to develop single-frequency precise point positioning algorithm for Global Positioning System (GPS), and hope to achieve the purpose of real-time positioning. In order to optimize the estimates, we adopt the following model to eliminate measurement errors: the Ultra-Rapid ephemeris provided by International GNSS Service (IGS) was used to correct satellite clock errors and calculate satellite position; besides, ionospheric errors was corrected by the Global Ionospheric Map which is also provided by IGS; tropospheric error was corrected by UNB3m model. In order to have a more stable and reliable positioning results, we combined Carrier smooth code (CSC) pseudorange and Doppler shift to estimate position and velocity and an adaptive Kalman filter was designed. The residual adaptive estimation was used to determine the state disturbance covariance matrix and big date analysis was performed to obtain the measurement noise covariance matrix. Moreover, this study attempts to perform the refinement of the extended Kalman filter. The applications of different algorithms to real experimental data show that the adaption method both on process disturbance and measurement noise covariance matrix can enhance the performance of positioning, and the refinement did well in the dynamic experiment. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/20541 |
DOI: | 10.6342/NTU201703177 |
全文授權: | 未授權 |
顯示於系所單位: | 應用力學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-106-1.pdf 目前未授權公開取用 | 2.59 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。