【摘要】：基于真實的GPS軌跡數(shù)據(jù),對城市擁堵路段進行預測.在此過程中,摒棄傳統(tǒng)的基于交通流預測和擁堵識別的方法,提出一種新的基于擁堵向量和擁堵轉移矩陣的擁堵路段預測方法.該方法同時考慮路段擁堵的時間周期性和時空相關性,通過對出租車GPS軌跡數(shù)據(jù)進行挖掘和訓練,建立擁堵向量和擁堵轉移矩陣,實現(xiàn)對擁堵路段的預測.真實數(shù)據(jù)集上的實驗驗證了所提的擁堵路段預測方法的有效性.
[Abstract]:Based on the real GPS trajectory data, the congested sections of the city are predicted. In this process, a new method based on congestion vector and congestion transfer matrix is proposed, which abandons the traditional methods based on traffic flow prediction and congestion identification. At the same time, the time periodicity and temporal and spatial correlation of congestion are considered in this method. By mining and training the GPS trajectory data of taxi, the congestion vector and congestion transfer matrix are established to predict the congested road section. Experiments on real data sets verify the effectiveness of the proposed method for predicting congested sections.
【作者單位】： 東北大學信息科學與工程學院;
【基金】：國家自然科學基金資助項目(61272177)
【分類號】：U491;TP311.13

【參考文獻】

相關期刊論文 前1條

1 姜桂艷;Q，

本文編號：2473411