טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentUrinovsky Suzanna
SubjectAccuracy Improvement in Existing Maps Using GPS Collected
Data
DepartmentDepartment of Civil and Environmental Engineering
Supervisors Professor Joshua S. Greenfeld
Professor Emeritus Yerach Doytsher
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Abstract

This research presents a process that has been developed to allow quick and low-cost correction of existing transportation network maps using points sampled by GPS navigation devices. The correction is performed in the area of sampling, as well as in area influenced by these corrections where no GPS points were sampled. The described research assumes that GPS navigation devices provide reliable and exact data, that will be used to correct maps with lesser accuracy compared to the GPS navigation devices. The large amount of data collected by navigation devices and their unique features allow utilization of these data in order to handle inaccuracies of positioning on map. Errors of tens of meters can be corrected in certain circumstances due to large amount of data and its distribution. The process that was developed in this research makes it possible to improve data bases of roads network quickly and with low costs, utilizing data derived from GPS navigation device. The process presents an innovative approach since it deals with GPS data as a source of information for correction of maps. Unlike this research, most of the works in this area deal with matching GPS data to maps and the improving of algorithm for positioning of GPS navigation devices on roads. Large quantities of GPS data is available for low cost in comparison to data collected by other means. Therefore, map correction using the proposed method makes it possible to correct data quickly, to receive statistical estimation of network accuracy improvement, to correct data on extensive area, to handle different infrastructure features such as electricity, gas, water networks etc. in addition to roads. The described research finds and defines several rules that will lead to the improvement of the original network data accuracy. Assorted algorithm quality verifications have been made during the research and conclusions have been made stating that the level of improvement of the original network accuracy varies between 25%-50%. It depends on the quality of the source road network data and of the GPS data (level of match between the network and the GPS points, distribution of GPS data sources, amount of sampled data, etc.). Furthermore, the level of improvement of the original network depends on following the GPS data sources distribution rules that have been defined in the research, in area statistical analysis and other factors.