| Popis: |
In view of dynamic changes of the road network, the aim of this essay is to recognize the location of the road with changed restriction information. Based on the FCD (floating car data), it studies the trajectory characteristics of the vehicle at first, including when the vehicle turning left, right or around. And then, it extracts the target points by using two criteria, one is the deflection angle of the vehicle, the other one is the change of the relative position. In addition, also the most important step, it determines the initial clustering center according to the density distribution characteristics of target points. Lastly, after clustered by K-means, it can identify quickly that the road allowed to turn or not. And if the traffic information of this road is same as the actual road, the restriction state is unchanged, otherwise, changed. Compared with the traditional algorithms, the accuracy of the optimized one is higher, and it can take advantage of FCD to implement the dynamic update of change information of the urban road network. |
