以L1單頻GPS進行個人移動定位

Abstract

利用全球導航衛星系統(GNSS, Global Navigation Satellite System)進行個人移動定位時，當GNSS接收儀對空通視有遮蔽物阻隔，常有訊號中斷的情況；現今行動裝置上常具備多種感測器與晶片，若以行動裝置上之動作感測器進行定位，則不受遮蔽影響。本研究以整合GNSS訊號與動作感測器的方式，探討提升定位服務之可靠度與不同GNSS定位技術成果在整合後之精度，以期建立高精度與可靠度的個人行動定位。 本研究課題有二，一為以各種GNSS定位設備與技術，探討GNSS可達之定位精度與可靠度之差異為何。二為整合動作感測器後，探討可達之定位精度及可靠度與整合前有何不同。本文中亦介紹研究使用之定位技術，包含做為精度參考之後處理定位方式，以及做為行動裝置外加GNSS定位之即時動態差分定位。實驗成果依整合動作感測器與否分為兩部分探討。整合前之GNSS成果，包含行動定位精度參考與即時動態定位成果；整合後之成果中，可依整合GNSS資料來源，分成使用行動裝置GNSS或外加GNSS進行整合。外加GNSS進行整合有nRTK與即時電碼差分兩種定位方式；動作感測器亦有多部不同行動裝置進行整合。 實驗成果顯示整合動作感測器後，評估坐標成果之標準差皆有所下降，以行動裝置GNSS定位成果整合動作感測器後，標準差最多可下降10m，而外接GNSS整合後，標準差則下降0.01m~0.05m。

GNSS (Global Navigation Satellite System) provides an efficient tool for the positioning of a person during movement. When the GNSS signal is blocked by obstacle such as buildings and trees, GNSS alone will not be able to determine the position. There are multi type sensors on a mobile phone unit currently available on market. Some of these sensors could be applied for positioning and not affected by the obstacles. This study integrates GNSS together with the motion sensors and the performance for positioning during movement is investigated. There are two subjects in this study. The performance of different GNSS solution schemes is examined first in terms of accuracy and reliability. Then, the improvement of integration is evaluated. The GNSS solution schemes evaluated include the rigorous post processing scheme which serves as a reference, and several real time kinematic positioning schemes such as network RTK, and real time code differential. The experiments are reported in two parts, without and with motion sensors. Several models of mobile devices are applied in the experiments for understanding the performance difference between models. The experiments show that the uncertainty of positioning is reduced when the motion sensor is applied. When the GNSS device on the mobile devices are used, the standard deviation of positioning error could be reduced by 10m. When the surveying type GNSS receiver is used, the standard deviation reduced between 0.01m and 0.05m in different cases.