Search

Ramp metering that uses traffic signals to regulate vehicle flows from the on-ramps has been widely implemented to improve vehicle mobility of the freeway. Previous studies generally update signal timings in real-time based on predefined traffic measurements collected by point detectors, such as traffic volumes and occupancies. Comparing with point detectors, traffic cameras—which have been increasingly deployed on road networks—could cover larger areas and provide more detailed traffic information. In this work, we propose a deep reinforcement learning (DRL) method to explore the potential of traffic video data in improving the efficiency of ramp metering. Vehicle locations are extracted from the traffic video frames and are reformed as position matrices. The proposed method takes the preprocessed video data as inputs and learns the optimal control strategies directly from the high-dimensional inputs. A series of simulation experiments based on real-world traffic data are conducted to evaluate the proposed approach. The results demonstrate that, in comparison with a state-of-the-practice method, the proposed DRL method results in (1) lower travel times in the mainline, (2) shorter vehicle queues at the on-ramp, and (3) higher traffic flows downstream of the merging area. The results suggest that the proposed method is able to extract useful information from the video data for better ramp metering controls. [ABSTRACT FROM AUTHOR]

Traffic delay is an effective index for estimating the performance of a signalized intersection. In this study, we provide a comprehensive review of the theoretical delay estimation model over the last ca. 90 years. For fixed-time signalized intersections, we classified the estimation development process into three stages. Stage 1 covered 1920s–1970s, when approaches based on steady-state theory were derived. These methods obtain accurate predictions with low degrees of saturation, but overestimate the delay with higher saturation and cannot provide reasonable results for oversaturated conditions. To accommodate high saturation, time-dependent models were proposed and improved in Stage 2, 1970s–2000s, using coordination transformation techniques. Progression factors to account for the filtering impact from upstream intersections were also introduced during this period. Due to inaccurate approximation of certain specific traffic conditions, some modified approaches and supplementary terms were derived from 2000 onwards (Stage 3), which facilitate the evolution of the delay estimation method and improved approximation results. Some new techniques, including artificial intelligence algorithms, were also introduced into delay estimation in this era. We also describe theoretical delay measurement methods for actuated control intersections with a similar time line. From our summary of the evolution of theoretical delay models, we highlight some deficiencies and future research directions. [ABSTRACT FROM PUBLISHER]

As transit agencies increasingly adopt the use of Automatic Passenger Count (APC) technologies, a large amount of boarding and alighting data are being amassed on an ongoing basis. These datasets offer opportunities to infer good estimates of passenger origin–destination (OD) flows. In this study, a method is proposed to estimate transit route passenger OD flow matrices for time-of-day periods based on OD flow information derived from labor-intensive onboard surveys and the large quantities of APC data that are becoming available. The computational feasibility of the proposed method is established and its accuracy is empirically evaluated using differences between the estimated OD flows and ground-truth observations on an operational bus route. To interpret the empirical differences from the ground-truth estimates, differences are also computed when using the state-of-the-practice Iterative Proportional Fitting (IPF) method to estimate the OD flows. The empirical results show that when using sufficient quantities of boarding and alighting data that can be readily obtained from APC-equipped buses, the estimates determined by the proposed method are better than those determined by the IPF method when no or a small sample sized onboard OD flow survey dataset is available and of similar quality to those determined by the IPF method when a large sample sized onboard OD flow survey dataset is available. Therefore, the proposed method offers the opportunity to forgo conducting costly onboard surveys for the purpose of OD flow estimation. [ABSTRACT FROM AUTHOR]