Dynamic Integration of Heterogeneous Transportation Modes under Disruptive Events
Dynamic Integration of Heterogeneous Transportation Modes under Disruptive Events
复制标题
DOI：
10.1109/iccps.2018.00015
复制
发表时间：
2018-04
期刊：
影响因子：
--
通讯作者：
Yukun Yuan;Desheng Zhang;Fei Miao;J. Stankovic;T. He;George Pappas;Shan Lin
中科院分区：
文献类型：
--
作者：
Yukun Yuan;Desheng Zhang;Fei Miao;J. Stankovic;T. He;George Pappas;Shan Lin
一个综合的城市交通系统通常由多种交通方式组成，这些交通方式在容量、速度和成本方面具有互补性，便于乘客根据计划时间表顺利换乘。然而，这样的集成并不被设计为在破坏性事件下操作，例如，地铁站的信号故障或公共汽车的故障，这些都会对乘客需求产生连锁反应，并大大增加延误。为了解决这些破坏性事件，当前的解决方案主要依赖于替代服务，以使用临时时间表和静态路线将乘客从受影响的区域运送到受影响的区域，例如，在关闭的地铁站发送穿梭巴士。这些解决方案效率极低，并且不利用实时数据来估计动态乘客需求。为了充分利用异构运输系统下的破坏性事件，我们设计了一个服务称为eRoute的分层滚动时域控制框架的基础上，自动重新路由，重新安排，并重新分配多模式的运输系统的基础上，实时和预测的需求和供应。我们着眼于地铁和公交的整合，使用大型数据集实现和评估eRoute，这些数据集包括（i）拥有13，000辆公交车的公交系统，（ii）拥有127个地铁站的地铁系统，（iii）拥有16，840个读卡器和800万卡用户的自动售检票系统。数据驱动的评估结果表明，与现有解决方案相比，我们的解决方案将服务乘客比率（RSP）提高了11.5倍，并将平均旅行时间减少了82.1%。
An integrated urban transportation system usually consists of multiple transport modes that have complementary characteristics of capacities, speeds, and costs, facilitating smooth passenger transfers according to planned schedules. However, such an integration is not designed to operate under disruptive events, e.g., a signal failure at a subway station or a breakdown of a bus, which have rippling effects on passenger demand and significantly increase delays. To address these disruptive events, current solutions mainly rely on a substitute service to transport passengers from and to affected areas using ad-hoc schedules and static routes, e.g., sending shuttles to closed subway stations. These solutions are highly inefficient and do not utilize real-time data to estimate dynamic passenger demand. To fully utilize heterogeneous transportation systems under disruptive events, we design a service called eRoute based on a hierarchical receding horizon control framework to automatically reroute, reschedule, and reallocate multi-mode transportation systems based on real-time and predicted demand and supply. Focusing on an integration of subway and bus, we implement and evaluate eRoute with large datasets including (i) a bus system with 13,000 buses, (ii) a subway system with 127 subway stations, (iii) an automatic fare collection system with a total of 16,840 readers and 8 million card users from a metropolitan city. The data-driven evaluation results show that our solution improves the ratio of served passengers (RSP) by up to 11.5 times and reduces the average traveling time by up to 82.1% compared with existing solutions.