A new partial ambiguity resolution method based on modified solution separation and GNSS epoch-differencing

摘要

Abstract Partial ambiguity resolution has been widely used in real-time kinematic (RTK) positioning to provide accurate and continuous centimeter-level positioning solutions. Current PAR methods are concerned with the exclusion of ambiguities that otherwise would result in low rate of full ambiguity resolution (FAR). With an assumption of unbiased ambiguities, the traditional PAR methods use the ambiguity-domain test statistics and probability-domain quantitative evaluation to select and validate the ambiguity subset. Consequentially, they would degrade the performance using low-cost devices in challenging environments where ambiguity biases exist, causing high probabilities of false alarm and missed detection of PAR and subsequently poor availability and accuracy of PAR. To deal with this issue, in this study, we propose a new PAR method for application in challenging environments. The proposed method consists of two major steps. First, a global navigation satellite system epoch-differencing (GED) algorithm is applied to derive a prior ambiguity solution. Second, we use a modified solution separation (SS) method with the prior ambiguity solution as an external input to obtain more accurate ambiguity test statistics. Based on a dynamic road test under environments with significant signal blockages, the performance of the proposed method is analyzed by using a low-cost GNSS receiver. The proposed method provides ambiguity test statistics with higher accuracy and can achieve 73.19 and 50.55 improvement in the accuracy and availability of the fixed solution, compared with the traditional PAR methods. Besides, the RMS of positioning errors with fixed solution are 1.03?cm, 0.70?cm, and 1.50?cm for the proposed PAR method in the east, north, and upward directions, respectively, which are 1.06?cm, 0.72?cm, and 1.35?cm for SS-based PAR, and 5.29?cm, 0.86?cm, and 5.56?cm for ILS-based PAR. The proposed PAR method achieves 90.36 fixed epochs, versus 79.06, and 88.28, for ILS-based PAR and SS-based PAR, respectively.Highlights Global navigation satellite system (GNSS) epoch differencing algorithm (GED) is used to derive a prior ambiguity solution.A modified solution separation (SS) method with the prior ambiguity solution as an external input is used to obtain more accurate ambiguity test statistics.We proposed a new partial ambiguity resolution (PAR) method based on SS and GED in realtime kinematic (RTK) positioning.A dynamic road test with a low-cost GNSS receiver shows that our method outperforms the significant improvements in both the accuracy and availability of RTK fixed solutions.