基於稀疏GPS探針車之幹道速度估測

Abstract

提供用路人各路段的即時交通資訊，並配合導航系統能有效的節省行車時間及能源消耗。由於具備GPS定位系統的車子逐漸普及，利用裝載GPS收發器的車子作為探針車取得即時交通資訊，並且回報給交通資訊中心來評估出目前的交通狀態成為可行的方式。 一般來說，交通資訊中心都是將回報的資料直接採用平均及歷史統計的方式來估測即時的交通狀態。但在少量回報資料時，會因為紅綠燈對探針車有不同的影響，而導致無法準確地估測交通狀況。本研究基於少量回報的情況下，利用走停模型及停走模型來降低紅綠燈所造成的影響並估測交通狀況。此外我們以蜂窩浮動車數據 (Cellular Floating Vehicle Data, CFVD) 為參考並比較所估測的旅行時間，在長期平均下相關係數約0.7而在即時估測下相關係數為0.5，顯示我們的方法與CFVD間擁有高度相似的交通估測。也因此我們結合兩者來提供可靠且高涵蓋的交通估測。

Providing real-time traffic information to road users and navigation systems can save travel time and reduce fuel consumption. With the increasing popularity of vehicles equipped with GPS receivers and wireless communication capability, the vehicles can be used as probes to collect the real-time traffic information. Since the probes report traffic information to a traffic information center (TIC), it is feasible to estimate real-time traffic condition. In general, the TIC uses the mean of the reported data from probes, at times considering historic statistics, to estimate real-time traffic conditions. However, when the probing data are sparse, the mean value may not be able to estimate the traffic conditions of surface roads accurately because traffic signals may cause random delays on probe vehicles. In this paper, we proposed Stop-Go and Go-Stop models to analyze probes’ GPS trajectory to reduce the influence of traffic light and estimate real-time traffic speeds with sparsely reported probe data. In addition, we compare our method with cellular floating vehicle data (CFVD). The correlation coefficient of out method and CFVD is about 0.7 in long time average and correlation coefficient is about 0.5 in real-time estimation. The results indicate that the traffic estimation of our method and CFVD are consistent. Therefore, combining both of them would provide reliable traffic estimation with a high coverage of the road network.