| Summary: | 碩士 === 國立交通大學 === 運輸與物流管理學系 === 106 === The mixed-traffic condition is common in most of the developing countries. In Taiwan, the large ratio of scooters in the mixed traffic increases the complexity of interactions between vehicles. To understand the capacity of signalised intersections, the discharge characteristics under the mixed-traffic condition is critical. The less bias on the estimation of discharge time can improve the estimation of discharge flow rate of intersections. The aim of this study is to understand the discharge behaviour of vehicles under the mixed-flow traffic condition and formulate a model to estimate the discharge time.
Past research discovered the relationship between the discharge time and the affecting factors, such as numbers of the vehicles, length of the queue, geometry design and layout, and the interactions between vehicles. However, there were limited studies that investigated the effect of order and arrangement of vehicles in a queue to the discharge process. In Taiwan’s Highway Capacity Manual, it is assumed that scooters concentrate at the scooter-waiting zone and their discharge do not affect the vehicles behind. However, in observations, scooters and other vehicles may mix up in the queue. This study proposed a new factor, Queue Pattern Entropy (QPE), which can describe the vehicle stopping sequence and queue pattern formed by different vehicle compositions.
Microscopic vehicle trajectory data collected from Unmanned Aerial Vehicle (UAV) is used in this study to gain insights into the interactions between vehicles. The dataset allowed us to observe traffic characteristics such as lateral movement, discharge order, and discharge times of the mixed traffic flow. Furthermore, a regression analysis is proposed to construct a discharge time estimation model under mixed-traffic condition. Linear and non-linear structure has been calibrated and compared to the model without QPE. Both forms of the model shows that the QPE is beneficial and superior to the base model. Compared with previous models in the literature, our model with QPE can effectively describe the queue pattern and attain less bias on the discharge time estimation.
|
|---|
