基於行動通信網路訊號之平面道路即時交通資訊

Abstract

近年來，隨著科技的進步，手機的普及率越來越高，有鑑於此，利用手機換手(Handover)的訊號作為定位的訊號，並且利用無線通訊網路追蹤使用者位置，透過多個手機即時換手訊號，推估目標路段即時交通資訊，這樣的機制不需要花費巨大的費用來建設或維護傳統的車輛偵測裝置，這種裝置損壞率及建設費用都相當高昂，並且只能提供固定地點的交通資訊。相較之下，透過無線通訊網路訊號追蹤分析使用者的移動行為，可以提供更即時、全面的交通資訊，而且額外設備的建設或維護費用大大降低。 本研究提出利用多組手機的換手訊號，這樣的研究，近年來越來越多，但是主要成果多著重在高速公路、快速道路等長途封閉式的道路，對於市區道路的研究較少，本研究針對利用無線通訊網路提供市區道路的即時交通資訊，提出一個可行的方法，首先透過觀察無線通訊網路控制訊號，找出目標路段上相關基地台(CELL)，觀察這些基地台的無線通訊網路控制訊號，挑選二個相同手機產生的合適換手訊號，組成換手對(Handover Pair)，針對每一個換手對收集相對應的旅行時間(Travel Time)，將這些旅行時間製作成累積分布函數(Cumulative distribution function, CDF)，透過CDF將旅行時間轉化為百分位的概念，透過百分位融合不同換手對的資訊產生交通狀況指數(Traffic Condition Index, TCI)，將TCI轉換成塞車等級(Congestion Level)，藉此提供用路人參考道路交通狀況。

As traffic congestion occurs more frequently in urban area, providing real-time traffic information can help road users to avoid congested areas and plan a better route. Estimating real-time traffic situations from the control signals, such as handovers, of the cellular network may be a cost-effective approach, compared with the conventional vehicle detectors and GPS-equipped probe vehicles, because no additional on-road or on-vehicle devices are needed. This approach has been referred to as Cellular Floating Vehicle Data (CFVD). However, CFVD have only been shown to be effective for freeway, but not for surface roads in urban area because of the smaller traffic flow and inaccurate positioning of mobile phones from the control signals. In this thesis, we present a novel CFVD method to estimate the traffic conditions of surface roads. First, we collect the travel time of each handoff pair for a long period time to establish Cumulative distribution function (CDF) of the travel time. Then, the travel time can be converted to a percentile by referring to the corresponding CDF. The percentiles of the travel times of different handover pairs in the same road are average to generate Traffic Condition Index (TCI), which can be used to determine the traffic congestion level. Using the anonymized handover records of both UMTS and GSM, we can obtain the historic traffic data that matches the daily traffic patterns on our experiment road segments. Assume that the historic traffic data is the pseudo ground truth; we have evaluated the hit rates of congestion levels. The results indicate that the real-time TCIs appear not to be reliable enough for estimating real-time traffic conditions. Further studies are needed to obtain more reliable estimation of the real-time traffic conditions of surface roads using CFVD.