Multi-criteria Genetic Algorithm-based Map-matching Method for Floating Car Data

To reduce the error and missing of the low-sampling-rate floating car data, a map-matching method based on multi-criteria genetic algorithm (MGA-MM) was established in this paper. The fitness function of the multi-criteria genetic algorithm contained the geometric similarity, distance of the shortest path and direction similarity. The dynamic time warping technique was introduced to estimate geometric similarity between recorded trajectory and observed route, and the distance of the shortest path was calculated by A^* algorithm. The data of low-sampling-rate floating car in Haidian District of Beijing was selected to conduct the corresponding test. Result shows that MGA-MM has ideal matching accuracy and fast running speed. When the interval of the sample is 10-20 s, the matching accuracy reaches 93.7%, which can satisfy the timeliness and accuracy of the map-matching of low-sampling-rate floating car data in traffic engineering.