Intracity heavy truck freight trips are basic data in city freight system planning and management. In the big data era, massive heavy truck GPS trajectories can be acquired cost effectively in real-time. Identifying freight trip ends (origins and destinations) from heavy truck GPS trajectories is an outstanding problem. Although previous studies proposed a variety of trip end identification methods from different perspectives, these studies subjectively defined key threshold parameters and ignored the complex intracity heavy truck travel characteristics. Here, we propose a data-driven trip end identification method in which the speed threshold for identifying truck stops and the multilevel time thresholds for distinguishing temporary stops and freight trip ends are objectively defined. Moreover, an appropriate time threshold level is dynamically selected by considering the intracity activity patterns of heavy trucks. Furthermore, we use urban road networks and point-of-interest (POI) data to eliminate misidentified trip ends to improve method accuracy. The validation results show that the accuracy of the method we propose is 87.45%. Our method incorporates the impact of the city freight context on truck trajectory characteristics, and its results can reflect the spatial distribution and chain patterns of intracity heavy truck freight trips, which have a wide range of practical applications.

中文翻译：

---

从重型卡车 GPS 轨迹识别同城货运行程结束

同城重卡货运行程是城市货运系统规划和管理的基础数据。在大数据时代，海量重卡 GPS 轨迹可以实时获取，性价比高。从重型卡车 GPS 轨迹识别货运终点（起点和目的地）是一个突出的问题。虽然以往的研究从不同的角度提出了多种行程终点识别方法，但这些研究主观地定义了关键阈值参数，忽略了复杂的城内重型卡车行驶特征。在这里，我们提出了一种数据驱动的行程终点识别方法，其中客观地定义了识别卡车停靠点的速度阈值和区分临时停靠点和货运行程终点的多级时间阈值。而且，通过考虑重型卡车的同城活动模式，动态选择合适的时间阈值水平。此外，我们使用城市道路网络和兴趣点 (POI) 数据来消除错误识别的行程终点，以提高方法的准确性。验证结果表明，我们提出的方法的准确率为87.45%。我们的方法结合了城市货运环境对卡车轨迹特征的影响，其结果可以反映城内重型卡车货运行程的空间分布和链条模式，具有广泛的实际应用。验证结果表明，我们提出的方法的准确率为87.45%。我们的方法结合了城市货运环境对卡车轨迹特征的影响，其结果可以反映城内重型卡车货运行程的空间分布和链条模式，具有广泛的实际应用。验证结果表明，我们提出的方法的准确率为87.45%。我们的方法结合了城市货运环境对卡车轨迹特征的影响，其结果可以反映城内重型卡车货运行程的空间分布和链条模式，具有广泛的实际应用。