Massive GNSS Network Analysis Without Baselines: Undifferenced Ambiguity Resolution

• Published in: Journal of geophysical research. Solid earth Vol. 126; no. 10
• Main Authors: Geng, Jianghui, Mao, Shuyin
• Format: Journal Article
• Language: English
• Published: 01.10.2021
• Subjects: ambiguity resolution; double‐difference; GNSS; network analysis; undifferenced
• ISSN: 2169-9313; 2169-9356
• DOI: 10.1029/2020JB021558

Integer ambiguity resolution (IAR) plays a key role in high‐precision Global Navigation Satellite System (GNSS). While GNSS network analysis is usually achieved through double‐difference (DD) IAR, undifferenced (UD) IAR has also been developed in recent years to achieve network analysis station by station. It is normally acquiesced in that DD‐IAR and UD‐IAR should achieve equivalent results, which is however not the case in a number of studies. We investigated four representatives DD‐IAR and UD‐IAR strategies using 1 year of GPS data from 192 globally distributed stations, where the position repeatability and the root mean squares error (RMSE) against the International GNSS Service weekly solutions were both quantified. We found that the position repeatability and the RMSE of DD‐IAR were both worse than those of UD‐IAR at over 90% of stations for the east component, though the largest deteriorations were only 0.6 and 0.7 mm, respectively. We demonstrate that it is the incorrectly resolved DD ambiguities, though accounting for far less than 1%, that degrade DD‐IAR. Identifying such problematic ambiguities requires sophisticated quality control for DD‐IAR, because the integer offset of an incorrect DD ambiguity might have been absorbed evenly by the cross‐connected mass of DD ambiguities throughout the network, and consequently does not manifest as an outlier in the ambiguity‐fixed solution. We, therefore, recommend UD‐IAR for GNSS network analysis since it is immune to the error propagation resulting from incorrect integer ambiguities at other stations and thus more efficient to achieve ambiguity‐fixed solutions. Plain Language Summary Integer ambiguity resolution (IAR), that is, fixing the unknown full cycles of carrier signal waves to integers, is significant to Global Navigation Satellite System precise positioning. There are two types of IAR, that is, double‐difference (DD) and undifferenced IAR. DD IAR is achieved by differencing between satellite pairs and station pairs, whereas undifferenced IAR is achieved without differencing observations between stations. In theory, the two IAR strategies can achieve equivalent results, but some studies showed that undifferenced IAR can achieve more precise positions than DD IAR. We demonstrate that it is the DD ambiguities fixed to incorrect integers that degrade the positioning precision. To identify all those incorrect ambiguities requires sophisticated quality control. In contrast, undifferenced IAR is far less affected by incorrect ambiguities. Besides, when processing a massive network, DD IAR is quite slow in computation. Therefore, we recommend undifferenced IAR since it is more resistant to incorrect ambiguity resolution and more efficient in achieving high‐precision positioning. Key Points Conventional Global Navigation Satellite System (GNSS) network analysis is subject to robust double‐difference ambiguity resolution Undifferenced ambiguity resolution is more robust in quality control and hence more efficient in massive network analysis PRIDE precise point positioning ‐AR has been open‐sourced with undifferenced ambiguity resolution capability for GNSS network analysis