A Real-Time Map-Matching Algorithm

Map matching is the problem of matching recorded geographic coordinates to a logical model of the real world. A GPS navigation device sends coordinates by an error between 2 and 10 meters. This causes confusion about the true location of the vehicle especially in dense urban areas. Matching a vehicle’s location to the true road segment is a challenge.
In this project, I have developed a real time mapmatching algorithm that can be used in navigational systems. Project is supported by Infotech company which develops gps systems to follow cars on traffic. My algorithm uses direction of the car as different from point to curve matching algorithm. I have tested the algorithm on digital map of Kadikoy, Istanbul. Compare to other geometrical approaches to mapmatching problem, the algorithm decreases error rate dramatically with a small increase in time complexity.