Air quality monitoring is essential to assess ambient air quality and its impact on commuters' activities and health. The fixed monitoring stations are cost-intensive, laborious, and provide limited data, while mobile monitoring using portable, low-cost air quality monitoring devices can provide high-resolution data. This study proposes a probabilistic approach to select the transit vehicles from the scheduled fleet, on which portable devices can be mounted to collect the spatiotemporal data points and maximize coverage continuously. General Transit Feed Specification (GTFS) and real-world vehicle data are employed. The proposed framework creates the segments, evaluates the overlaps, and prioritizes the routes based on joint probabilities. The bipolar Sigmoid activation function is used for probability estimation. A minimum number of portable devices installed on selected vehicles can provide maximum spatiotemporal coverage with the least overlapping routes for real-time air quality monitoring in larger cities. One of the major limitations of the framework is that it requires a schedule of transport modes on which portable devices are to be installed. The results and inferences can be beneficial to the relevant authorities for the appropriate installation of portable, low-cost air quality monitoring devices.

中文翻译：

---

实时空气质量监测的路线选择，最大化时空覆盖

空气质量监测对于评估环境空气质量及其对通勤者活动和健康的影响至关重要。固定监测站成本高昂、费力且提供的数据有限，而使用便携式、低成本空气质量监测设备的移动监测可以提供高分辨率数据。本研究提出了一种从预定车队中选择交通车辆的概率方法，在这些车辆上可以安装便携式设备来收集时空数据点并持续最大化覆盖范围。采用通用运输供给规范 (GTFS) 和实际车辆数据。所提出的框架创建分段，评估重叠，并根据联合概率对路线进行优先级排序。双极 Sigmoid 激活函数用于概率估计。在选定的车辆上安装最少数量的便携式设备，可以以最少的重叠路线提供最大的时空覆盖，从而实现大城市的实时空气质量监测。该框架的主要限制之一是它需要安装便携式设备的运输模式时间表。结果和推论可能有利于有关当局适当安装便携式、低成本的空气质量监测设备。