Easing Traffic Congestion: An Improved Clustering Method for Sharing Bike Station Deployment

An excessive number of cars cause serious traffic jams. Fortunately, a new kind of environmentally friendly transportation service, sharing bikes, came into being. In the cities with shared bikes, deploying shared bikes stations purposefully will make a contribution to reducing the pressure of the traffic. We aim to draw support from sharing bikes to improve the bad traffic. To find the real problems of the current traffic. We make full use of history taxi trajectories to analyze current traffic condition. We design a traffic jam detection framework in this paper. It is called CF framework for short. Derived from the density-based clustering algorithm of inspiration, we propose a new clustering method (CF-Dbscan). The new method has successfully been applied to the trajectories clustering. To deal with errors of devices, a road network matching algorithm (CF-Matching) helps match GPS points to real road network accurately. The first experiment proves that our clustering algorithm performs better than DBSCAN in the field of trajectory clustering. We design another experiment to verify the effectiveness of our CF framework in the real scene. The results of the experiments prove that we can achieve the purpose of reducing traffic jam with our framework.