Investigation of Virtual RINEX data quality

Network Real Time Kinematic (NRTK) Global Navigation Satellite System (GNSS) technology is being utilised for a wide range of positioning applications. The advantage of NRTK is its ability to provide corrections (accounting for atmospheric and satellite orbit errors) that are based on a Continuously Operating Reference Station (CORS) network rather than a single reference station. While most users employ NRTK for real-time applications, it is also possible to benefit from network-based GNSS corrections for post-processing applications. This is achieved through the provision of Virtual RINEX data, i.e. data that would have been observed at an imaginary, unoccupied (i.e. virtual) GNSS reference station whose location is specified by the user. This paper presents an initial investigation into the quality of Virtual RINEX data in regards to processing outcomes. An extensive 3-day dataset is used to compare static positioning results obtained with Virtual RINEX data generated by CORSnet-NSW and RINEX data observed at two test sites (incorporating small and large NRTK cells) in New South Wales, Australia. At each test site, data are analysed in three ways: (1) ‘zero’ baseline processing between virtual and observed data for session lengths ranging from 10 minutes to 24 hours, (2) AUSPOS processing using virtual and observed data for 2-hour, 6-hour and 24-hour sessions, and (3) baseline processing relative to surrounding CORS using virtual and observed data for session lengths of 10 minutes, 1 hour and 24 hours. It is found that ‘zero’ baselines vary from 1 mm (hz) and 2 mm (vt) for long observation sessions in a small NRTK cell to 15 mm (hz) and 40 mm (vt) for all observation windows investigated in a large NRTK cell. 24-hour AUSPOS solutions based on Virtual RINEX data agree with those using observed data at the 10 mm level or better, while 2-hour solutions show differences of up to about 20 mm (hz) and 40 mm (vt). Baseline processing to surrounding CORS reveals differences ranging from the few-mm level for short (10 km) baselines in a small NRTK cell to the few-cm level for long (70 km) baselines in a large NRTK cell. These results indicate that Virtual RINEX data are comparable to observed data for some applications, provided NRTK cell size, observation length and baseline length are taken into consideration.