A new method to improve the performance of multi-GNSS pseudorange positioning in signal-degraded environment

Abstract There are two existing multi-GNSS positioning methods, (1) separate receiver clock parameter is set for each constellation, (2) estimating intersystem biases (ISBs) in advance to obtain position solutions with only four unknowns. The former is the regular method and its unknowns include three receiver-coordinate parameters and several receiver clock parameters (depend on the number of participant constellations), so that it may disable when few satellites belonging to different GNSS are in view. The latter is one workable way to obtain position solution with only four visible satellites. In addition to the disabled regular method, the positioning results by the ISB-corrected method are often unsatisfactory for navigation users in signal-degraded environment. Both the deviation of ISB-solutions and the remaining measurement errors of the need corrected observations are factors to degrade location precision. Apart from these, fewer visible satellites usually cause a low robustness of the positioning model, which cause the negative influence of various errors is amplified. Based on the variation of various measurement errors, we propose a new ISSB-corrected method with observations corrected by corresponding satellite-dependent parameters. The new parameter contains the difference of time scales, hardware delays and uncorrected measurement errors between the corresponding satellite and reference, in other words, it can synthetically consider measurement errors and the ISB. By the ISSB-corrected method, we not only achieve positioning solutions with four satellites, but also significantly reduce the accuracy loss. Many experiments are conducted to present the superiority of the ISSB-corrected method. In open-area, the accuracies of regular and ISB-corrected methods are nearly equal. Apart from a similar accuracy in horizontal, the accuracy is improved by approximate 10% in up direction with respect to the two existing methods. Given the high redundancy of model in open-area, the new method may not improve the accuracy remarkably. However, it can make great contributions in signal-degraded environments. In order to compare the performance of ISB- and ISSB-corrected methods in environment with limited visible satellites, we simulate several scenarios (different satellites participant or various receivers) with only four participant satellites in the positioning solution. By the ISSB-corrected method, the 3D RMS of positioning results with four satellites is about 15 m, while it is usually worse than 25 m for the ISB-corrected method. In an urban vehicular test, the horizontal positioning error of the ISSB-corrected method is less than 20 m; and the ISB-corrected method may reach up to 70 m.