The General Transit Feed Specification (GTFS) is an open standard format for recording transit information, utilized by thousands of transit agencies worldwide. In this study, a new tool named GTFS2STN for converting GTFS data into the spatiotemporal networks is introduced. To analyze the travel time variability, it is important to transform a transit network to a spatiotemporal network to enable a comprehensive analysis of transit system accessibility. GTFS2STN is a new tool that converts General Transit Feed Specification (GTFS) data into spatiotemporal networks, addressing the lack of open-source solutions for transit analysis. The tool includes a web application that generates isochrone maps and calculates travel time variability between locations. Validation against Google Maps APIs shows that journey time (i.e., the summation of the transit time, walking time, and waiting time) differences in the Mean Absolute Percentage Error are typically within 12%. A before–after analysis shows that for the transit journey time in 2024 in Nashville, Tennessee, 8 out of 10 pivotal bus stops showed a significantly decreased journey time compared with the case of 2019. A further set of before–after analyses shows that although journey time between transit sites significantly dropped on May 2020 during COVID-19 emergencies, the journey time almost totally recovered to the before-COVID-19 level by November 2020. By supporting any valid GTFS schedule, GTFS2STN enables the analysis of historical and planned transit systems, making it valuable for long-term accessibility assessment and travel time variability studies.

Monitoring large trucks (LT) plays an essential role in the operation and planning of transportation and economics. Although many “advanced” traffic detectors (e.g., infrared, microwave, and magnetic sensors) can identify LT, those technologies are not as widely deployed as loop detectors, and their performance is disputed. While many previous studies have investigated estimation of truck traffic using gross volume and occupancy information from loop detectors, most of them are based on mean effective vehicle length (MEVL), and thereby vulnerable to dynamic traffic conditions and dependence on exogenous data sources. This paper aims to develop an original method of estimating LT volume using detector output (i.e., volume and occupancy). A linear optimization problem is formulated by minimizing the difference between observed occupancy and theoretical “best guess.” LT and short vehicle (SV) lengths and their counts are devised as decision variables to compensate for variations in speed and length. To assess the proposed method, volume and occupancy data output in 30 s granularity from three different cases are prepared: microscopic simulation, Tennessee Radar Detector System (TN-RDS), and California Freeway Performance Measures System (PeMS). For the simulation, 12 different scenarios were built up and the results indicate that the mean absolute error of LT estimates increases with traffic volume and truck penetration. In two real-world cases, the LT estimates are close to ground truth from video and weigh-in-motion (WIM) station records. Furthermore, this simple and online LT volume estimate approach can be extended to other stations and to monitor a wide range of LT activities.

Ramp metering is the act of controlling on-going vehicles to the highway mainlines. Decades of practices of ramp metering have proved that ramp metering can decrease total travel time, mitigate shockwaves, decrease rear-end collisions by smoothing the traffic interweaving process, etc. Besides traditional control algorithm like ALINEA, Deep Reinforcement Learning (DRL) algorithms have been introduced to build a finer control. However, two remaining challenges still hinder DRL from being implemented in the real world: (1) some assumptions of algorithms are hard to be matched in the real world; (2) the rich input states may make the model vulnerable to attacks and data noises. To investigate these issues, we propose a Deep Q-Learning algorithm using only loop detectors information as inputs in this study. Then, a set of False Data Injection attacks and random noise attack are designed to investigate the robustness of the model. The major benefit of the model is that it can be applied to almost any ramp metering sites regardless of the road geometries and layouts. Besides outcompeting the ALINEA method, the Deep Q-Learning method also shows a good robustness through training among very different demands and geometries. For example, during the testing case in I-24 near Murfreesboro, TN, the model shows its robustness as it still outperforms ALINEA algorithm under Fast Gradient Sign Method attacks. Unlike many previous studies, the model is trained and tested in completely different environments to show the capabilities of the model.