機車於市區道路受公車影響之駕駛行為研究

Abstract

台灣私人運具的使用以機車最為普遍，機車在混合車流中沒有車道規範，且不受小汽車停等車隊的影響，只要行駛路線遇到空隙便可持續推進。機車通常行駛在較外側車道，而公車進出站勢必會占用到外側車道的空間。公車體積大、行駛速度不如機車高、機動性低，機車駕駛較不願意整個路段持續跟車在公車後，經常性的從左側或右側超車，因此機車於市區道路受公車的駕駛行為是值得探討的領域。一般車流模擬軟體開發通常依照國外之交通環境所建立，係以小汽車為主要車種，並不符合國內特殊的背景。因此若能將國內機車於市區道路受公車影響的模式應用至模擬軟體中，可提升軟體之使用性與擬真程度。 由於近幾年攝影技術進步，蒐集市區道路的車流影像，取得原始資料相對容易得多，本研究將利用無人機進行空中拍攝，包含晨峰與中午時段且具國內市區道路代表性的台北市路段，共兩個方向，分別為兩小時完整週期的影片，涵蓋車流密度高、中之車流影像，配合點取車流軌跡軟體平台的開發(Lee et al., 2008b)，建立車輛軌跡的資料庫。 本研究利用上述所建立的資料庫發展機車受公車駕駛影響的行為模式。研究方法將採用多項羅吉特模式，基於隨機效用最大化以及個體理性選擇的原則下構建模式，加入公車之駕駛特性，考慮公車在公車站範圍會有進站、公車站、停靠等狀態，並將公車狀態以虛擬變數呈現，假設機車有三種方案包含直行跟車、左偏、右偏可選擇，了解機車面臨前方公車不同狀態時的因應方式，建立三種情境下的機車行為模式。本研究將結果應用至車流模擬軟體，探討模擬結果是否更符合真實情況。未來若能透過模擬軟體的結果，延伸討論道路標線設計、是否有快慢車道劃分、公車站的設置方式、公車站於路段中的位置等，有助於市區道路交通工程面的設計，進一步有效降低市區道路車流延滯的狀況。

Motorcycle is one of the major transportation modes in Taiwan. It is an irreplaceable vehicle type with its high motility and convenience. As the rapid growth of the population and the quantity of motorcycles, more and more traffic problems are arising, especially on urban arterials. The conventional traffic theories were almost devoted to the traffic flow of passenger car, and studies about the characteristics and driving behavior of motorcycles was limited. However, the mixed traffic composed of buses, cars and motorcycles might be a complicated topic because of the erratic driving behavior of motorcyclists such as filtering, swerving, oblique following. The motorcyclists usually face the approaching bus when travelling on the curb lane near the bus stop. The driven route and the line of sight of motorcyclists are blocked by the buses which are characteristics with lower speed and larger volume. Therefore, the buses could cause a serious impact on the motorcyclists on urban arterials. Instead of following the buses conservatively, the motorcyclists would like to overtake the buses soon and leave it far away. For simplicity, the driving behavior of motorcycles affected by buses is worth discussing in depth; however, only little research has attempted to study the movement of motorcyclists in the past. Besides, the traffic simulation software was mainly developed in European and American countries. The simulated results are unable to present the properties of traffic in Taiwan realistically without considering the mixed traffic flow and non-lane-based movements. We collected traffic data using aerial videography by a multicopter, which could be used to capture images and videos of the road traffic. The data was collected on a selected site location in Taipei. After extracting the trajectories and types of vehicles from the video images, the database could be built for analysis. The goal of this study is to formulate the driving behavior of motorcycles which is affected by buses as a discrete choice model using multinomial logit model, considering the driving state of buses simultaneously. The staggered car following of motorcycles is also discussed in this study. It is found that the driving characteristics of motorcycles when following the buses have totally difference between on the road and near the bus stop. According to the results of proposed models, the value of parameters could be explained reasonably and represent the behavior of motorcyclists correctly. The calibrated models are implemented into a traffic simulation model, Bikesim, to examine the performances of the models as compared to the real data. Our simulation results confirm that the proposed approach could reproduce the movements of motorcycles near a bus stop.