Transportation Research Record: Journal of the Transportation Research Board

Findings from previous studies indicate that a freeway traffic collision is more likely to occur in close physical proximity to the tail of a queue. The implication is that collision likelihood increases when drivers abruptly alter their trajectories (e.g., by decelerating or changing lanes) on encountering the queue. The implication is supported and bolstered with new and detailed data that were painstakingly extracted from two freeway stretches in California. These data show how the likelihood of collision increases as both the spatial and the temporal proximities to the tail of an expanding or receding queue become smaller. It follows that collision risk may be reduced by instructing drivers to begin decelerating while still upstream of queues. Retarding vehicle progress toward a queue's tail could retard the rate by which the likelihood of collision grows with time. Having vehicles approach a queue at diminished speeds may diminish the expected severity of a collision, should one still occur. This information is considered in this development of the ideas behind simple control logic for issuing speed advisories to drivers as they approach a queue. This logic could assist freeway managers in selecting suitable advisories so as to diminish the empirical estimate of collision likelihood at a specified time in the immediate future and by some target amount. In the interim, the likelihood estimates would be based solely on vehicles’ spatiotemporal proximities to queues. Other influences would be ignored for now, although planned experiments to test and improve the present logic are discussed.

Ramp metering (RM) is the most effective dynamic traffic measure in response to growing congestion in urban freeway networks. Among the extensive RM methods available, those based on optimal control theory have shown strong potential in improving freeway performance. However, these algorithms require an accurate traffic model that limits their applicability in practice. Reinforcement learning (RL) provides the tools to achieve optimal RM control without reliance on any traffic model. In this paper, a guideline for designing RM control systems based on RL is presented by testing different states’ representations, learning methods, action selection, and reward definitions. A microscopic simulation test bed based on a portion of Highway 401 in Toronto, Canada, is developed to evaluate each of the above design parameters and quantify various RM control strategies. A comparison of the reinforcement learning ramp-metering (RLRM) algorithm with a modified version of ALINEA shows the potential of RLRM to improve freeway traffic conditions. When applied to the developed case study, the proposed RLRM algorithm and modified ALINEA reduce the total travel time by 40% and 20%, respectively, compared with the case with no RM.

Travel time reliability is considered to be one of the key indicators of transport system performance. Knowledge of the mechanisms of travel time unreliability enables the derivation of explanatory models with which travel time reliability can be predicted and utilized in traffic management. Inspired by the macroscopic fundamental diagram (MFD), describing the relationship between production (average flow completing the trips) and vehicle accumulation (average density) in a traffic network, this paper investigates a macroscopic travel time reliability diagram (MRD) relating the travel time reliability to the network accumulation. The MFD relationship has potential because it characterizes the state of an entire traffic network with just two (production and accumulation) or three (adding spatial variability of accumulation) state variables. Likewise, the MRD describes the network travel time reliability as a function of just one independent state variable (network accumulation). Empirical analyses were performed to investigate the variability in MFD as seen in scatters and to show the travel time reliability in relation to the network accumulations. Traffic data from Dutch freeway networks were employed to facilitate the analyses. With the MRD for different freeway networks, it was found that a critical travel time reliability accumulation existed, below which network accumulation had little or no impact on travel time reliability and above which it had a significant impact. Also, the critical travel time reliability accumulation was, in general, lower than the critical MFD accumulation. These findings provide insights for road authorities in how to make trade-offs in traffic management between maximum production and travel time reliability.

The significant impact of heavy vehicles (HVs) on freeway operations has been identified since the first edition of the Highway Capacity Manual (HCM). The HCM 2010 used passenger car equivalent (PCE) values and percentages of trucks, buses, and recreational vehicles to account for the effect of HVs on roadway performance. Unfortunately, the PCE values in the HCM 2010 relied on a limited field database and simulation runs that were calibrated for steady-flow traffic conditions, although the effect of HVs on traffic flow could reasonably have been expected to have varied with traffic conditions. Few studies have been conducted with extensive field data to examine impacts on traffic characteristics by HV presence under congested and forced-flow conditions. This paper presents such an effort by using urban freeway data containing 1.2 million individual vehicle observations. The results indicated a significant difference in lagging–leading behavior between vehicle pairs related to HV presence. Passenger car and HV headways were found to increase with HV presence in the traffic stream. A similar pattern was found for the PCE factor. The PCE value under congested conditions and more than 9% HV presence was found to be 1.76, which was higher than the 1.5 value recommended by the HCM 2010 for level freeway sections. The maximum throughput of the freeway was found to be affected by HV presence. The maximum throughput was observed at a truck presence of 3%, after which it became progressively lower. The results of this paper can be used as input for future simulation runs of congested freeway flow conditions.

The U.S. Department of Transportation affirmed the integrated corridor management (ICM) initiative to mitigate traffic congestion on urban corridor networks by systematically leveraging and diverting traffic to better utilize the available capacities of parallel arterials. ICM strategies would be of utmost importance when traffic incidents that may cause significant delays occur on freeways. By effectively diverting upcoming traffic to the adjacent arterials via variable message signs and real-time traveler information systems (e.g., mobile devices), the impact of incidents can be alleviated. To understand fully the effectiveness of ICM strategies, a framework needs to be developed to model, simulate, and analyze ICM operations under different scenarios. This study developed and calibrated a VISSIM simulation model based on field-collected video and sensor data for the South of Downtown area of Seattle, Washington. The analysis aimed to quantify networkwide ICM performance by empirically diverting traffic to adjacent arterials in response to incident management for freeway operations. Multiple scenarios were built into the simulation model to account for different diversion rates as well as recurrent and nonrecurrent congestion situations. When more traffic is diverted into an arterial, the arterial's performance degrades while improvements in travel time and delay for the freeway become marginal. Through quantitative analysis of the delay, throughput, and travel time of freeway segments and the arterial network, the trade-off in the overall system performance was carefully studied and understood. The research findings set up a solid foundation for ICM strategy development and optimization of traffic system operations.

The primary objective of this study is to develop a robust queue estimation algorithm for motorway on-ramps. Real-time queue information is a vital input for dynamic queue management on metered on-ramps. Accurate and reliable queue information enables the management of on-ramp queues in a manner that adapts to the actual traffic queue size and thus minimizes the adverse impacts of queue flush while increasing the benefit of ramp metering. The proposed algorithm is based on the Kalman filter framework. The fundamental conservation model is used to estimate the system state (queue size) with the flow-in and flow-out measurements. These projection results are updated with the measurement equation by using the time occupancies from midlink and link entrance loop detectors. This study also proposes a novel singular-point correction method. This method resets the estimated system state to eliminate the counting errors that accumulate over time. In the performance evaluation, the proposed algorithm demonstrated accurate and reliable performance and consistently outperformed the benchmarked single-occupancy Kalman filter (SOKF) method. The improvements over the SOKF method were 62% and 63% on average for the estimation accuracy and reliability, respectively. The benefit of the innovative concepts of the algorithm is well justified by the improved estimation performance in conditions of congested ramp traffic, in which long queues may significantly compromise the benchmark algorithm's performance.

A number of intelligent transportation system technologies are deployed for disseminating information to travelers. One of the most important parameters for assessing the impacts and benefits of these deployments that influence route selection is the diversion rates of travelers. In this research, a method was developed to estimate traffic diversions on the basis of main-line detector data without the need for off-ramp detectors. In addition, the estimation allows the assessment of the guidelines and procedures for disseminating information and evaluating the impacts on alternative routes. The validity of the developed methodology was verified by comparing the estimated values with real-world data. Case studies of the developed method indicated that the average diversion rate was about 10% to 35% for three-lane and four-lane freeways, depending on the number of lanes blocked. A linear relationship between the average diversion rate and the lane blockage ratio was also developed; this relationship indicated that there was a general trend of increase in diversion with the increase in the lane blockage ratio.

This paper presents a methodology for incorporating freeway reliability analysis in the Highway Capacity Manual on the basis of Strategic Highway Research Program Project L08. The methodology uses a scenario-based approach in which each scenario represents a unique combination of traffic demands and facility (segment) capacities. Demand impacts consider variability in time-of-day, day-of-week, and month-of-year differences, and capacity impacts quantify effects of various nonrecurring congestion sources, such as weather, incidents, work zones, and special events. The method combines three key components: a data repository, a freeway scenario generator (FSG), and a computational software engine, FREEVAL-RL. The FSG determines the number of scenarios for the reliability analysis on the basis of the facility-specific combination of demand, incident, and weather events. For each scenario, it produces a set of inputs and adjustment factors that are passed on to the FREEVAL-RL engine, along with each scenario's probability. The engine is capable of batch-processing many scenarios and estimates the reliability distribution for the facility. At the core, the computational engine and underlying methodology are consistent with the freeway facilities method in the Highway Capacity Manual 2010, which is capable of estimating undersaturated and congested flow conditions in a multi-period analysis. The method is illustrated with a real-world case study freeway facility in North Carolina. The FSG for that facility resulted in 2,508 scenarios. The resulting travel time distribution is presented, and sensitivity analyses are presented to explore the contributing effects of weather and incidents on the overall reliability of the facility.

This paper examines a way to synthesize route travel time probability density functions (PDFs) on the basis of segment-level PDFs. Real-world data from I-5 in Sacramento, California, are employed. The first finding is that careful filtering is required to extract useful travel times from the raw data because trip times, not travel times, are observed (i.e., the movement of vehicles between locations). The second finding is that significant correlations exist between individual vehicle travel times for adjacent segments. Two analyses are done in this regard: one predicts downstream travel times on the basis of upstream travel times, and the second checks for correlations in travel times between upstream and downstream segments. The results of these analyses suggest that strong positive correlations exist. The third finding is that comonotonicity, or perfect positive dependence, can be assumed when route travel time PDFs are generated from segment PDFs. Kolmogorov–Smirnov tests show that travel times synthesized from the segment-specific data are statistically different only under highly congested conditions, and even then, the percentage differences in the distributions of the synthesized and actual travel times are small. The fourth finding, somewhat tangential, is that there is little variation in individual driver travel times under given operating conditions. This is an important finding, because such an assumption serves as the basis for all traffic simulation models.

Nonrecurring traffic incidents, such as motor vehicle crashes, increase not only travel delays but also the risk of secondary crashes. Secondary crashes can cause additional traffic delays and reduce safety. Implementation of effective countermeasures to prevent or reduce secondary crashes requires that their characteristics be investigated. However, the related research has been limited, largely because of the lack of detailed incident and traffic data necessary to identify secondary crashes. Existing approaches, such as static methods employed to identify secondary crashes, cannot fully capture potential secondary crashes because of fixed spatiotemporal identification criteria. Improved approaches are needed to categorize secondary crashes accurately for further analysis. This paper develops an enhanced approach for identifying secondary crashes that uses the existing crash database and archived traffic data from highway sensors. The proposed method is threefold: (a) defining secondary crashes, (b) examining the impact range of primary crashes that possibly relate to secondary crashes, and (c) identifying secondary crashes. The proposed methodology establishes a practical framework for mining secondary crashes from existing sensor data and crash records. A case study was performed on a 27-mi segment of a major highway in New Jersey to illustrate the performance of the proposed approach. The results show that the proposed method provides a more reliable and efficient categorization of secondary crashes than commonly used approaches.

This study presents the development and application of methodologies to support weather-responsive traffic management (WRTM) strategies by building on traffic estimation and prediction system models. First, a systematic framework for implementing and evaluating WRTM strategies under severe weather conditions is developed. This framework includes activities for planning, preparing, and deploying WRTM strategies in three different time frames: long-term strategic planning, short-term tactical planning, and real-time traffic management center operations. Next, the evaluation of various strategies is demonstrated with locally calibrated network simulation-assignment model capabilities, and special-purpose key performance indicators are introduced. Three types of WRTM strategies [demand management, advisory and control variable message signs (VMSs), and incident management VMSs] are applied to multiple major U.S. areas, namely, Chicago, Illinois; Salt Lake City, Utah; and the Long Island area in New York. The analysis results illustrate the benefits of WRTM under inclement weather conditions and emphasize the importance of incorporating a predictive capability into selecting and deploying WRTM strategies.

In 2008 the City of Orlando, Florida, embarked on construction of a new Amway Center to replace the existing Amway Arena. The new center was designed to host a variety of events such as concerts, arena football, graduation ceremonies, National Collegiate Athletic Association basketball, and the National Basketball Association's (NBA's) Orlando Magic home games. The Amway Center, which opened in October 2010, had a capacity of 19,000 patrons for NBA games. The Amway Center was adjacent to Interstate 4 and State Road 408 in the downtown core. To prepare for the Amway Center's event management challenges, the City of Orlando's Transportation Engineering Department utilized smart event traffic management techniques, which integrated the following system and plan components: intelligent transportation system infrastructure (software and hardware), traffic control device upgrades, regional planning efforts, and state-of-the-art safety and security protocols. These components also helped minimize the environmental impacts of traffic congestion. When designing smart event plans, it is important to consider high-threshold events such as the NBA All-Star Game for system performance and to understand the balance required for operational efficiency and security in today's global climate. In addition, infrastructure performance is critical during high-profile events so that the local economy can benefit and the municipality can stay competitive to attract more of these types of events. Finally, the resultant operational efficiency is vital for reducing the carbon footprint and negative societal impacts.

This paper describes the impacts on transit performance from three separate conversions of high-occupancy vehicle lanes into variably priced high-occupancy toll (HOT) lanes in Miami, Florida; Minneapolis, Minnesota; and Atlanta, Georgia. The data showed that the conversions had no negative impact on bus performance. In all three cities, the buses experienced travel time savings after the conversion: 17 min in Miami, 4.5 min in Minneapolis, and 5 min in Atlanta. Similarly, the HOT conversions did not negatively affect bus ridership. Ridership increased over the baseline by 57% in Miami and 13% in Minneapolis. In Atlanta, ridership increased by 11%, and this increase began before the conversion. Bus riders’ overall perception of the HOT lanes has been positive. In surveys, the bus riders from Miami and Minneapolis gave high ratings for travel times and reliability in the HOT lane corridors. Riders in Atlanta rated the bus service as very good but were not as positive about the tolls. Riders disagreed that the HOT conversion had improved their travel or been good for the Atlanta region. These negative responses may be attributable in part to widespread disgruntlement with the HOT concept that existed even before implementation.

Microscopic simulation is critical for evaluating the operation strategies of managed lanes. However, most existing tools are limited in their ability to simulate dynamic tolling strategies of managed lanes, particularly those with multiple segments. Three sets of modeling components are developed in this paper to demonstrate simulation of high-occupancy toll (HOT) lane operations. The first component implements three pricing strategies: responsive pricing; a closed-loop, control-based algorithm; and time-of-day pricing. The second component mimics drivers’ lane choice behaviors in the presence of tolls, and the third represents different toll structures for multisegment HOT lane facilities. An enhanced version of CORSIM, which contains these new modeling components, is validated by simulation experiments involving the 95 Express network in South Florida.

In response to an impending demand growth at an existing reactive managed lane system and to provide a timely and more effective temporary hard shoulder activation, this paper presents the development of short-term prediction models. A lane-oriented attribute—namely, the left-lane flow distribution ratio—is introduced to ameliorate the system by capturing the forthcoming stream dynamics and reconfiguring it to be proactive. To assess the impact on the network's performance of implementing a system for hard shoulder running, an exploratory analysis was performed on the basis of data acquired by seven radar sensors located every 500 m along a Swiss freeway section that was not affected by incoming or exiting traffic. A locally weighted regression was employed to provide more accurate insight into traffic behavior by comparing observations derived during the regular operation of the system and a period in which it was suspended, with respect to seasonality patterns. To describe the impending stream motion by examining different time–volume clusters (off-peak and peak hours), two prediction models were specified according to the time range. The preliminary results of the study for several prediction horizons demonstrate an acceptable prediction uncertainty. The hard shoulder activation prediction confirms the analysis of the findings of this research with regard to the impact on operations.

Lane changing is one important aspect of vehicle maneuvers that has significant impacts on freeway operation. Most of the previous research on lane changing has focused on the microscopic modeling of lane change movements. Very few studies have investigated lane-changing behavior along high-occupancy vehicle (HOV) facilities, especially in the context of comparing different access types. This paper describes work that extracted lane change data with high resolution in both time and space from videos and investigated lane-changing behavior on freeways with different types of HOV access configurations. Furthermore, the extracted lane change data were correlated with lane-by-lane flow and occupancy data from the California Freeway Performance Measurement System. Compared with continuous-access HOV facilities, limited-access HOV facilities were found to have a higher number of lane changes per unit distance and a shorter time gap when the subject vehicle moved out of the HOV lane. Also, most of the lane changes on limited-access HOV facilities occurred within the first half of ingress–egress areas. The insights provided by these results on lane-changing maneuvers along HOV facilities can be used to support traffic operation analyses, calibrate driver behavioral logic in traffic simulation models, and inform the design of lane configurations and access control measures for managed lanes.