FASF法和最小二乘寻优法用于途中模糊值解算的比较

A Comparison of the FAST and Least-Squares Search Algorithmus for Ambiquity On The Fly

介绍FASF法的理论、算法和陆海空验算结果,并同目前常用的最小二乘寻优法进行比较。FASF法的重要特征是,确定每个模糊值的寻优范围。过去,每个模糊值的寻优范围是分别按其它模糊值的假定整数来确定,而FASF法的寻优范围则是按递归法确定。它在确定模糊值的范围时,顾及了假定模糊值对其它模糊值的影响。另一特征是,利用卡尔曼滤波顾及了从初始历元到现今历元的全部观测值。因此,无需寻找一切可能的模糊值,从而使计算工作量急剧减小。

This paper describes a new and effective approach for ambiguity resolution on the fly (OTF), namely the Fast Ambiguity Search filter (FASF), and presents comparative numerical results between this method and the least-squares search approach. One of the critical aspects of the FASF approach is the determination of the search range for each ambiguity parameter. In the past, the search range of each ambiguity was determined separately regardless of the assumed intergers of the other ambiquities. With FASF, the search ranges are determined recursively and are releted to each other. To determine the uncertainty range of an ambiguity parameter, the effect of an assumed integer on other ambiquities is fully taken into account. Another important characters tic is that all observations from the initial to the current epoch are taken into account through the use of Kalman filter. An index of possible inabilities to fix the ambiguites is used. Therefore, a full search of all possible integer ambiguities is not required and the computiion is dramatically reduced. Computer memory is also saved since there is no need to retain all potential solitions. Test results show that the computation time for ambiguity search is nearly negligible. Most important, there is a very significant redution in the observation period required to resolve the ambiguities OTF, as demonstrated using L1 and widelane land, areborne and marine data.