Accuracy and reliability of map-matched GPS coordinates: the dependence on terrain model resolution and interpolation algorithm

  • Authors:
  • Jing Li;George Taylor;David B. Kidner

  • Affiliations:
  • GIS Research Centre, School of Computing, University of Glamorgan, Pontypridd, Wales, CF37 1DL, UK;GIS Research Centre, School of Computing, University of Glamorgan, Pontypridd, Wales, CF37 1DL, UK;GIS Research Centre, School of Computing, University of Glamorgan, Pontypridd, Wales, CF37 1DL, UK

  • Venue:
  • Computers & Geosciences
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The global positioning system (GPS) has become the most extensively used positioning and navigation tool in the world. Applications of GPS abound in surveying, mapping, transportation, agriculture, military planning, GIS, and the geosciences. However, the positional and elevation accuracy of any given GPS location is prone to error, due to a number of factors. This has serious implications for some applications, such as real-time navigational systems. GPS accuracy can be significantly improved with additional data, possibly from multiple sources, and especially from multiple receivers. In the case of a single GPS receiver, its position and elevation can be considerably improved with the use of spatial data. For vehicle tracking, map matching can be employed to intelligently snap the GPS location to a road centreline, while height aiding can augment the GPS solution by utilising a digital terrain model (DTM), thereby reducing the number of satellites required to determine a position. This paper describes the use of map matching and height aiding, and examines the effect of different terrain resolutions (Ordnance Survey 1:50,000 and 1:10,000 scale DTMs) on plan position and elevation accuracy for vehicle tracking. Furthermore, the user's choice of interpolation algorithm for estimating heights from the DTM is investigated. The results of the experiments described in this paper demonstrate that height aiding alone reduces the mean error in elevation from 22.5 to 17.5m for of a single GPS receiver, and the mean error in plan position from 6 to 5m. However, map matching and height aiding combined, reduces the elevation RMSE of a single GPS receiver from 22.5m to approximately 4m (1:50,000 scale DTM) and down to 0.8m (1:10,000 scale DTM), while the plan position RMSE is reduced from 5.9 to 3.2m (either DTM). It is also demonstrated that when the number of satellites visible to the receiver is reduced, or the satellite geometry is poor, map matching and height aiding considerably improves the plan and elevation accuracy. The use of a higher-order interpolant (e.g. a bicubic or biquintic polynomial) is shown to slightly improve performance, compared to a bilinear interpolant, for the lower-resolution DTM, but has little overall benefit for the higher resolution DTM.