Low-cost accuracy for ITS applications from a national GNSS network
By Martti Kirkko-Jaakkola, Stefan Söderholm, Salomon Honkala, Hannu Koivula, Sonja Nyberg, and Heidi Kuusniemi, Finnish Geospatial Research Institute (FGI), National Land Survey of Finland
Our real-time kinematic (RTK) implementation, the Public Precise Positioning (P3-Service) project, has achieved horizontal positioning accuracy of 0.5 meters using relatively inexpensive equipment: a commercial off-the-shelf (COTS) low-cost GNSS receiver. The project used FinnRef, the Finnish national GNSS network.
With inter-station baselines on the order of 200 kilometers, FinnRef is relatively sparse in comparison with commercial RTK networks. We used FinnRef as the RTK base station, either in single-base or network RTK mode. Although FinnRef’s main purpose is to maintain the national coordinate system, it is also capable of delivering DGNSS and network RTK data over the NTRIP protocol.
Transport Applications. Horizontal position accuracy of 0.5 meters or better, achieved for more than 90 percent of the time with small, low-cost devices, could be useful in various applications, particularly in intelligent transportation systems.
Current consumer-grade GNSS solutions routinely offer a positioning accuracy in the order of 5 meters, and satellite-based augmentation systems (SBAS) such as WAAS and EGNOS can improve the accuracy to the order of 1 meter. However, this is not adequate for all use cases; in particular, intelligent transportation systems (ITS) require better positioning performance. For instance, a horizontal accuracy of 0.5 meters is needed to reliably identify the lane in which a vehicle is driving. Maintaining inventory of machines, road signs and other infrastructure would also benefit from sub-meter accuracy.
Sub-meter or even sub-decimeter positioning accuracies can be attained with a relatively good reliability in real time if a dual-frequency GNSS receiver and a physical or virtual base station are available. However, such receivers and virtual base station services are currently too expensive to gain popularity in the mass market. Recently, precise point positioning (PPP) has demonstrated that comparable accuracies can be attained without a base station using real-time precise correction data, but its drawback is a long convergence time. In contrast, differential methods utilizing raw base-station observations, such as RTK, converge much faster.
Network RTK Test. Network RTK performance was tested in a static scenario with the closest physical base station 63 kilometers from the rover receiver. Network corrections were delivered in the PRS representation, and data were logged for 20 minutes at a rate of 1 Hz. The plot above shows the resulting horizontal position errors. The dashed red circle with a radius of 0.5 meters centered at the reference location (green triangle) contains 90.4 percent of the position estimates.
For a full account of the experiments and results described here, see the paper “Low-Cost Precise Posioning Using a National GNSS Network,” presented at ION GNSS+ 2015.
Follow Us