號誌時制設計適用之機車小客車當量值分析

Abstract

混合車流之道路環境為台灣及多數東南亞國家所特有之交通特色，其中又以機車車輛為多數居民主要之交通工具，機車與小客車之駕駛行為特性具有多處不同，如機車不似小客車遵從車道行駛，而會鑽行於車道與車輛之間。為了行車安全及行車效率之考量，國內針對機車車輛之駕駛行為特性設立了特有之道路設計，而此些道路設計形成了機車車輛特有之疏解特性，因此造就了與多數東南亞國家不同之混合車流道路環境。 在道路設計之實務上通常以小客車當量值（Passenger Car Equivalent）將異質車流轉換為同質車流，過去研究中多以駕駛行為特性之因素來估計小客車車當量值，如行車速度、佔有面積與間距等等，然而鮮少以號誌時制設計為考量進行機車小客車當量值之估計。 適當之號誌時制設計可以有效減少交叉口之衝突與延滯，並且提高交叉口之績效與駕駛人之行車效率，而在號誌時制設計中交通流量之多寡恆為重要考量之一，因此機車小客車當量值之估計對於號誌時制設計之影響相當重要，若能求得適當之機車小客車當量值，則可準確估計各方向或各流動之綠燈時間需求，以此設計出最適當之號誌時制。 本研究蒐集不同車流情境之停等車隊疏解時間等資料，以此建立不同車流情境之綠燈時間需求模型，並推估基於綠燈時間需求模型之機車小客車當量值。此外，本研究分別以簡單線性迴歸模型、假設資料服從指數分佈之非線性迴歸模型與二次迴歸模型分析機車車輛數與小客車車輛數之關係，由結果顯示大部分之車流情境較適用於非線性迴歸模型與二次迴歸模型，表示機車車輛數與小客車車輛數之關係應受到不同交通組成而有所變動，以此推論機車小客車當量值應為一變動值。本研究另以最適當之配適迴歸模型求得之機車小客車當量值估計值建立機車小客車當量值推估模式，藉由此模式根據交叉口各流動之車流情境估計適用之機車小客車當量值，並將該機車小客車當量值納入後續號誌時制設計中。 本研究以臨界流動設計方法之概念，研擬最適週期臨界流動重整方法與最適綠燈時間需求臨界流動重整方法，最適週期之臨界流動重整方法以Webster-Cobbe公式求得最適週期長度，並以臨界流動作為分配綠燈時間之依據，而最適綠燈時間需求之臨界流動重整方法則以綠燈時間需求模型求得各流動所需之綠燈時間需求，亦以臨界流動作為分配綠燈時間之依據。求得各個重整之號誌時制後，利用VISSIM微觀車流模擬軟體建立道路環境，並以平均停等延滯與飽和程度做為績效之評估指標。 由結果可知，以適用於不同交叉口車流情境建立之機車小客車當量值推估模式求得之機車小客車當量值除了能降低整體交叉口之停等延滯，亦能滿足交叉口各個方向與各個流動之綠燈時間需求。因此，本研究認為在設計號誌時制時應將停等車隊之疏解時間納入考量，始能展現機車停等區得以在短時間內疏解機車之特性，此外，本研究認為應根據不同車流情境給予相對應之機車小客車當量值，故根據所得之最適機車小客車當量值給予在不同混合車道總疏解時間與混合比之車流情境下所適用之機車小客車當量值建議值，當混合車道總疏解時間大於50秒且混合比大於50%時之機車小客車當量值應介於0.00-0.10，當混合車道總疏解時間大於50秒且混合比小於50%時之機車小客車當量值應介於0.10-0.15，當混合車道總疏解時間介於30秒至50秒且混合比大於30%時之機車小客車當量值應介於0.15-0.20，當混合車道總疏解時間介於30秒至50秒且混合比小於30%時之機車小客車當量值應介於0.20-0.30，當混合車道總疏解時間小於30秒時之機車小客車當量值應介於0.20-0.30。

Mixed traffic flow is the specific characteristic for Taiwan and some Southeast Asian countries. Besides, the scooter is the main transportation mode in these countries. The driving behavior of the scooter is quite different from that of the car. For example, the scooter will filter between lanes and vehicles rather than follow the lane discipline. In the consideration of the driving safety and driving efficiency, there are specific road designs for the driving behavior of the scooter in domestic. The road designs cause the specific departing property of the scooter and also create the mixed traffic flow different from some Southeast Asian countries for Taiwan. It is common that the heterogeneous traffic flow is converted into homogeneous traffic flow by Passenger Car Equivalent (PCE). In the past, the factors which can indicate the driving behavior are used to estimate Passenger Car Equivalent of scooter which are traffic speed, occupied space, headway, and so on. However, it is few to estimate the Passenger Car Equivalent of scooter considering the signal timing design. The appropriate signal timing can not only reduce the conflict and delay in the intersection effectively, but also improve the performance of the intersection and the driving efficiency. Among the signal timing design, traffic flow will be one of the vital considerations. Therefore, the estimation of Passenger Car Equivalent of scooter is quite important to the signal timing. The right Passenger Car Equivalent of scooter can bring the correct green-split demand for each direction or each movement and design the most appropriate signal timing. In this study, the data of the departure time for the stop queue is collected for constructing the Green-Split Demand Model (G-D model) under different traffic conditions and estimating the Passenger Car Equivalent of scooter based on Green-Split Demand Model. Moreover, the different Passenger Car Equivalent of scooter by the different traffic conditions are considered in this study. Besides, there are three kinds of regression models which are the simple linear regression model, the nonlinear regression model with the exponential distribution, and the quadratic regression model. It is found that the most of traffic conditions are more suitable to the nonlinear regression models and the quadratic regression models which means that the relationship between the amount of scooters and cars should vary by the different traffic conditions. It is concluded that the Passenger Car Equivalent of scooter should be a varied value. In this study, the estimated models for the Passenger Car Equivalent are constructed by the estimated values of Passenger Car Equivalent from the most appropriate regression model. By the estimated models, the Passenger Car Equivalent of scooter can be estimated based on different traffic conditions and those values are applied to the successive design for the signal timing. In this study, the critical movement signal re-timing method based on the cycle length and the critical movement signal re-timing method based on the green-split demand are planned based on the concept of Critical Movement Method. The critical movement signal re-timing method based on the cycle length calculates the most appropriate cycle length by Webster-Cobbe Formula and distributes the green-split by the critical movement. The critical movement signal re-timing method based on the green-split demand calculates the green-split demand required by each movement by Green-Split Demand Model and also distributes the green-split by the critical movement. The microscopic traffic simulation software, VISSIM, is used to construct the road environment with the signal re-timing programs and the average stop delay and the degree of saturation are used to be the performance indicators. It is found that the signal re-timing method with the Passenger Car Equivalent of scooter estimated by the regression model based on each intersection can not only reduce the average stop delay of the intersection, but also satisfy the green-split demand of each direction and each movement. Therefore, the result indicates the considerations of the departure time of stop queue can show the property of the starting waiting zone which can depart substantial amounts of scooter in a short period. Besides, the varied Passenger Car Equivalent of scooter for different traffic condition is essential to the signal timing design. The suggested values of Passenger Car Equivalent are given in the traffic conditions with the different total departure time in the mixture lane and the different mixture rates. The value of Passenger Car Equivalent should be 0.00 to 0.10 when the total departure time in the mixture lane is lager than 50 seconds and the mixture rate is lager than 50%. The value of Passenger Car Equivalent should be 0.10 to 0.15 when the total departure time in the mixture lane is lager than 50 seconds and the mixture rate is smaller than 50%. The value of Passenger Car Equivalent should be 0.15 to 0.20 when the total departure time in the mixture lane is between 30 to 50 seconds and the mixture rate is larger than 30%. The value of Passenger Car Equivalent should be 0.20 to 0.30 when the total departure time in the mixture lane is between 30 to 50 seconds and the mixture rate is smaller than 30%. The value of Passenger Car Equivalent should be 0.20 to 0.30 when the total departure time in the mixture lane is smaller than 30 seconds.