Realignment of Road Network Maps with GPS Tracking Data
Nicola JAMMALIEH, Joshua GREENFELD
Professor Emeritus New Jersey Institute of Technology, and Faculty of Civil and Environmental Engineering
Road network datasets are widely available either for a fee or for free on the Internet.
Unfortunately, some of them are not always accurate and up-to-date. These inaccuracies could
cause navigation errors and prove costly to users. Therefore, it is important to devise a useful,
efficient and cost effective method to make the datasets more accurate. One way to rectify the
dataset quality is to use GPS data collected by GPS enabled navigation devices. When the
map is not accurate it is reasonable to assume that the GPS data is more accurate than the
map. Thus, GPS tracks can be used to realign the traveled street segments. One can view this
as the inverse of the map matching problem. Instead of matching GPS positions to the map,
we match the map to GPS tracks (or points).
This paper outlines a comprehensive approach for realigning street segments to GPS data
collected from moving vehicles. The process includes GPS data filtering, matching GPS
points to existing road segments, shifting the road segments to the GPS points and forming
new intersections and vertices. The end result of the process is a revised map of the road
segments in their corrected positions. For each of these tasks new algorithms or enhanced
existing algorithms were developed and employed. The proposed process was successfully
implemented on real world data and the results of the realigned road segments are shown,
analyzed and verified. The realigned network showed full agreement with high accuracy
orthophoto of the test area.
Event: FIG WW 2019 Hanoi
Only personal, non-commercial use of this document is allowed.