|M.Sc Student||Nicola Jammalieh|
|Subject||A Heuristic Method for Improving Digital Road Maps Using|
Kinematic GPS Data
|Department||Department of Civil and Environmental Engineering||Supervisors||Full Professor Greenfeld Joshua S.|
|Professor Emeritus Doytsher Yerach|
|Full Thesis text - in Hebrew|
There are now computerized road maps such as, Google Earth or maps acquired with navigation systems that cover a large percentage of Earth's surfaces. The accuracy of these maps is not always of the highest quality. The accuracy of the maps can be improved by using GPS data collected with various navigation devices. More accurate GPS points should be utilized during the process of anchoring the maps to the GPS data. Improving the accuracy of the maps can contribute to all sorts of areas such as the navigation (more precise spatial information of vehicle location relative to nearest intersection, along the route, and the like) and map matching (improving the performance of map matching algorithms). Today there are databases that keep the GPS traveled data collected with standard navigation equipment. This data are available to all users and can be obtained free of charge. One can use this data to enhance existing computerized maps.
In this document a method for enhancing the quality of maps using GPS point collected by moving vehicle is presented. There are several studies related to the improvement of the maps. These studies dealt with matching a single road segment to the GPS data. They suggested ways to resolve the problem by extracting nodes from GPS data instead of matching the road segments to the GPS data. The purpose of this study is to solve the problem globally, that is to match the road network data to the GPS data instead of matching them to only a single road section. In this thesis I present possible solution methods to match the map to GPS data.
After running the algorithm of improving the accuracy of the maps it is necessary to assess the resulting map (revised) compared to the original map. This is done by comparing the original and revised maps and to an orthophoto with an accuracy up to 0.5 m. Results showed a significantly improvement of the location of the intersection. Intersections in the revised map moved closer to their true location. The average distances between the intersection points of original map and orthophoto map improved by 11.20 meters after running the algorithm. This testifies to quality and effectiveness of the suggested solution.