RTK Ambiguity Resolution Validation Using a Genetic Algorithm Approach

Marco Mendonca, Altti Jokinen, Ryan Yang, Gary Hau, Yi-Fen Tseng

Abstract: In the context of Real-Time Kinematic (RTK) positioning, the vast majority of applications expect, or even require, the precision level provided by the fixed solution.The required fixed ambiguity terms can be estimated trivially, being the step known as the "validation," the one with challenging aspects. Once the ambiguity terms are estimated, the validation step attempts to verify whether the set of ambiguities is likely to be corresponding to the real number of integer carrier-phase cycles from satellite to receiver’s antenna. Upon failing to discriminate a good solution from a poor solution, the filter state estimation becomes susceptible to wrong fixes, deteriorating the integrity of the estimated position. In this context, this paper approaches an improvement of the ratio test ambiguity validation by the lenses of a Genetic Algorithm (GA) machine learning classifier. A large dataset of kinematic data in urban environments were utilized to train the GA function to discriminate "good" and "bad" fixes. This discrimination is performed by estimating an adaptive validation threshold and comparing the method with a ratio test solution for validating ambiguities. Four aspects are assessed: The receiver operating characteristic (ROC) curve, the percentage of fixed epochs, percentage of wrong fixes, and overall solution accuracy. The results show that the GA method can approximate the validation test to the global maximum of the likelihood function and classify wrong fixes as well as good fixes better than a traditional ratio test at its optimum performance. At its optimum classifying point, the GA method outperformed the traditional ratio test in all categories considered, showing improvements of 48% in accuracy, 71.73% in dispersion, and 24.15% in the 95th percentile of solutions. Considering that this method can be applied in real-time on a mass-market GNSS receiver, the achieved results show a promising novel method for many applications requiring low-cost, and reliable results.
Published in: Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022)
September 19 - 23, 2022
Hyatt Regency Denver
Denver, Colorado
Pages: 2701 - 2705
Cite this article: Mendonca, Marco, Jokinen, Altti, Yang, Ryan, Hau, Gary, Tseng, Yi-Fen, "RTK Ambiguity Resolution Validation Using a Genetic Algorithm Approach," Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022), Denver, Colorado, September 2022, pp. 2701-2705.
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In