自駕運具對於都市運具分配之影響分析

Abstract

自駕運具為近年蓬勃發展的先進運具，全球團隊致力開發打造能夠正式上路運行之自駕車，除了自用小客車及共享車隊之形式外，全球亦有多個自駕巴士測試運行之成功案例。對於自駕運具引入所帶來的運具轉移效應，國際上已有數篇相關研究，然而多數研究並未將機車納入考量。台灣的交通特性有相當高比例的機車使用，未來自駕運具引入以後使用者如何做運具選擇，是一值得研究課題。因此，本研究以台北都會區為例，採用多項羅吉特模式，考量自行車、機車、小客車、公車、捷運及自駕巴士等五項運具，分別建立使用者之效用函數，求得目標年2025年引入自駕巴士前後之各運具比例。此外並針對自駕巴士設定四種定價方式，以分析引入不同營運模式自駕巴士對於運具轉移之影響。 研究結果顯示，2025年引入自駕巴士前，台北都會區運具比例分別為：自行車7.96%, 機車26.30%、汽車18.36%、公車18.32%以及捷運29.06%。在引入單一票價、分段票價、距離收費及隨選服務之自駕巴士以後，台北都會區之自駕巴士運具市場分別有5.79%、2.19%、7.24%及10.63%，並且大多由機車、公車、捷運轉移而來。平均旅次長度部分，上述四種情境之平均旅次距離為5.04公里、4.81公里、5.22公里及4.93公里。此外，本研究對自駕巴士之效用函數，分別對旅行時間價值、等車時間以及票價進行敏感度分析，結果顯示自駕巴士之運具比例對自駕巴士速度、車外旅行時間、單一票價及單位距離票價較敏感，可以做為未來研擬營運與行銷策略之參考。 本研究建立之運具分配模型，針對自駕巴士營運情境以及四種定價方式引入後，分析都市運具比例之變化與衝擊，並且以台北都會區為案例進行分析。研究成果將有助於未來臺北都會以及其他地區進行自駕巴士服務之先期規劃。

The rapid development of autonomous vehicle has become main focus throughout the world, as it is regarded as an integration of advanced technologies that will bring major impacts to transportation sector. Aside from autonomous car models and related shared mobility, many driverless buses have successfully completed various trails in numerous areas. Recent trends in the autonomous vehicle development have led to a proliferation of studies on the user behavior and mode shift with implementation of autonomous vehicles. However, most studies in the field have not included motorcycles as an alternative, resulting in a lack of understanding for mode shift in areas having a high proportion of motorcycle usages, such as Taiwan. Therefore, this study applies the multinomial logit model to explore the effects of driverless bus on mode shift. Utility functions are defined for bike, motorcycle, car, bus, metro and 4 scenarios for driverless bus while Taipei Metropolitan in 2025 was analyzed as a case study. Research results have first verified the initial mode shares without the implementation of driverless bus are as follows: bike 7.96%, motorcycles 26.30%, cars 18.36%, bus 18.32% and metro 29.06%. The driverless bus mode share of the 4 scenarios, namely Flat Fare, Distance Based I, Distance Based II and On Demand Service are: 5.79%, 2.19%, 7.24% and 10.63%, respectively. Various average trip distances are also obtained for the 4 scenarios: 5.04km, 4.81km, 5.22km and 4.93 km, respectively. Sensitivity analysis has also conducted and it is shown that the driverless bus share is more sensitive to bus speed, out of vehicle travel time, flat fare and price charged per km. These findings may be considered for formulating strategies of promoting driverless bus in the future. The mode split model developed in this study contributes to the understanding of driverless bus induced mode shifts in high motorcycle usage areas, and provided numerical analysis for Taipei Metropolitan. The methodology can also be applied to other areas in order to investigate the effects of driverless bus on mode shift. The model developed and research results will be a reference for preliminary planning of driverless bus service in Taipei Metropolitan and other cities.