論文翻譯標題: 提升適用於移動製圖應用之多元感測器整合策略 與融合演算法
  • 楊 成中

學生論文: Doctoral Thesis


Mobile mapping systems (MMSs) have been widely used to rapidly acquire spatial information In such a system the integration of the Inertial Navigation System (INS) and the Global Positioning System (GPS) is popularly applied as a direct geo-referencing system to seamlessly determine the time-variable position and orientation parameters Given that both INS and GPS are affected by various error sources these errors deteriorate the overall position and orientation accuracy of an integrated system that uses conventional INS/GPS integration strategies These strategies include loosely coupled and common estimation strategies such as Kalman filter These errors particularly occur when using low cost small size inertial sensors and GPS receivers or operating in GPS-hostile environments such as in urban canyons The use of expensive equipment can improve the performance of the system but also increases its overall cost In some operating conditions the high cost devices still do not fulfill the user’s requirements with the current dada fusion techniques This study aims to improve the navigation solution accuracy of a land-based MMS utilizing a low cost inertial sensor and GPS receiver by focusing on integration strategies and advanced estimation algorithms For INS/GPS integration a modified tightly coupled scheme was proposed to overcome the limitations of current tightly coupled and loosely coupled schemes Various additional aids and constrains are discussed and applied to improve the robustness of the system in cases of GPS signal outages For the estimation strategies the efficiency of different filtering techniques such as KF Unscented KF particle filter and hybrid filtering are analyzed using simulation and real data A new estimation method called on-line smoothing is proposed to overcome the problems of filtering techniques and conventional smoothing algorithms Various field tests using different devices and testing scenarios were conducted to evaluate the proposed methods The analyzed results indicate that the proposed strategies and algorithms can significantly improve the performance of the system compared with conventional schemes In various testing scenarios the improvement of the modified TC over pure LC is about 60% and in GPS-denied environment with the aid of additional constrains the improvement of proposed scheme can reach to 90% over the scheme with INS/GPS
獎項日期2014 一月 20
監督員Kai-Wei Chiang (Supervisor)