對單點GPS 定位誤差進行時間序列分析，將定位誤差擬合成十階AR 模型。AR(10)模型的零極點分析結果表明模型的動態特性由一對主導極點決定，得到單點GPS 定位誤差的二階馬爾可夫模型，有效提高定位精度。模型階次的降低，減少了定位解算的復雜度和計算量，提高了系統的實時性。仿真結果顯示：將GPS 定位誤差擬合成二階馬爾可夫過程是合理的。
關鍵詞：單點GPS；定位誤差；建模；AR 模型；二階Markov 模型
Abstract: GPS positioning error is modeled to be the 10-order AR model on the basis of the time series analyses. Through the frequency domain analysis of AR(10), the system dynamic characteristics is determined by the dual main poles of AR(10) model, so the single-stand GPS positioning error can be described as the 2-order Markov process. The positioning precision is improved. The low- order model makes the positioning calculation easier and quicker. Through simulation, it is reasonable that GPS positioning error is modeled to be the 2-order Markov process.
Keywords: single-stand GPS; positioning error; modeling; AR model; 2-order Markov model