arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2207.11285physics.soc-ph
收藏

基于大规模网约车的城市道路网络最优路侧感知

Optimal Drive-by Sensing in Urban Road Networks with Large-scale Ridesourcing Vehicles

Shuocheng Guo, Xinwu Qian

AI总结:

针对传统城市道路感知方法成本高、覆盖有限等问题,本研究提出基于大规模网约车的最优路侧感知模型与启发式算法,在保障服务水平的同时显著提升感知覆盖度与可靠性,且成本低廉。

摘要:

城市道路网络的感知与监测有助于城市交通系统的高效运行和城市系统的功能发挥。然而,传统感知方法,如感应线圈传感器、路侧摄像头以及来自海量城市出行者的众包数据(例如Google Maps),往往受限于成本高、覆盖范围有限和可靠性低等问题。本研究探讨了路侧感知(drive-by sensing)的潜力,这是一种利用大规模网约车(ridesourcing vehicles, RVs)开展城市道路网络监测的创新方法。我们首先通过历史路段到访情况,从覆盖度和可靠性两个维度评估了网约车的感知性能。接下来,我们提出了一种基于行程的最优网约车重路由模型,在保障网约车出行服务水平不变的前提下,最大化感知覆盖度和可靠性。此外,在假设行程相互独立的情况下,我们设计了一种可扩展的基于列生成的启发式算法,用于规划网约车的巡航轨迹。我们利用纽约市超过20000辆网约车的真实轨迹数据,通过实验和敏感性分析验证了所提模型的有效性。优化后的重路由策略效果显著提升:与历史记录相比,道路网络的显式感知覆盖度提高了15.0%至17.3%(随时段变化),感知可靠性至少提升了24.6%。扩大路径搜索空间可进一步将感知覆盖度最多提升4.5%,可靠性提升超过4.2%。此外,考虑到网约车司机的激励因素,感知性能提升的成本极低,每位司机仅需0.10美元,凸显了其成本效益。

英文摘要:

The sensing and monitoring of the urban road network contribute to the efficient operation of the urban transportation system and the functionality of urban systems. However, traditional sensing methods, such as inductive loop sensors, roadside cameras, and crowdsourcing data from massive urban travelers (e.g., Google Maps), are often hindered by high costs, limited coverage, and low reliability. This study explores the potential of drive-by sensing, an innovative approach that employs large-scale ridesourcing vehicles (RVs) for urban road network monitoring. We first evaluate RV sensing performance by coverage and reliability through historical road segment visits. Next, we propose an optimal trip-based RV rerouting model to maximize the sensing coverage and reliability while preserving the same level of service for the RVs' mobility service. Furthermore, a scalable column generation-based heuristic is designed to guide the cruising trajectory of RVs, assuming trip independence. The effectiveness of the proposed model is validated through experiments and sensitivity analyses using real-world RV trajectory data of over 20,000 vehicles in New York City. The optimized rerouting strategy has yielded significantly improved results, elevating explicit sensing coverage of the road network by 15.0\% to 17.3\% (varies by time of day) and achieving an impressive enhancement in sensing reliability by at least 24.6\% compared to historical records. Expanding the path-searching space further improved sensing coverage of up to 4.5\% and reliability of over 4.2\%. Moreover, considering incentives for RV drivers, the enhanced sensing performance comes at a remarkably low cost of \$0.10 per RV driver, highlighting its cost-effectiveness.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑