利用GPS衛星訊號的觀測量進行即時動態定位，在有限的單時刻訊號觀測量數據與極短時間內要完成動態定位演算，最大的困難即是整數相位模稜(Phase Ambiguity)、幾何坐標分量與電離層延遲之間參數高相關的問題，造成整數相位模稜值非常不易正確求解。本論文研究即在探討如何使參數與觀測式間，組合最佳化之演算工作模式，並應用方差無偏差估計技術-- BIQUE估計式，估計方差分量，使訊號觀測量數據回饋提供最佳化無偏差之先驗資訊，讓隨機模式能更接近事實，以強化解關聯演算技術之工作效能，使整數相位模稜估計求解正確且更具效率。並引用卡曼濾波器進一步探討附加參數對幾何定位之影響與效益，期獲得最佳之精確定位成果。 Using GPS data for real-time kinematic differential positioning, the most difficulty is to solve high correlation between ambiguities, geometry and ionospheric parameters. Precise estimation of GPS phase ambiguity integer is a key. In this study, some possible enhancements which try to take advantage of the structure between the parameters and double-differenced observation equations in the data processing will be given, and take an algorithm of variance and covariance component unbiased estimation — BIQUE. Let the stochastic model close to truth, and fast solve integer phase ambiguity resolution while a decorrelation method is applied. This study will discuss how to increase the parameters to geometry position effectively when the Kalman filter is used.
