Lee D. Han

Online-SVR for short-term traffic flow prediction under typical and atypical traffic conditions

Most literature on short-term traffic flow forecasting focused mainly on normal, or non-incident, conditions and, hence, limited their applicability when traffic flow forecasting is most needed, i.e., incident and atypical conditions. Accurate prediction of short-term traffic flow under atypical conditions, such as vehicular crashes, inclement weather, work zone, and holidays, is crucial to effective and proactive traffic management systems in the context of intelligent transportation systems (ITS) and, more specifically, dynamic traffic assignment (DTA). To this end, this paper presents an application of a supervised statistical learning technique called Online Support Vector machine for Regression, or OL-SVR, for the prediction of short-term freeway traffic flow under both typical and atypical conditions. The OL-SVR model is compared with three well-known prediction models including Gaussian maximum likelihood (GML), Holt exponential smoothing, and artificial neural net models. The resultant performance comparisons suggest that GML, which relies heavily on the recurring characteristics of day-to-day traffic, performs slightly better than other models under typical traffic conditions, as demonstrated by previous studies. Yet OL-SVR is the best performer under non-recurring atypical traffic conditions. It appears that for deployed ITS systems that gear toward timely response to real-world atypical and incident situations, OL-SVR may be a better tool than GML.

Global Optimization of Emergency Evacuation Assignments

Conventional emergency evacuation plans often assign evacuees to fixed routes or destinations based mainly on geographic proximity. Such approaches can be inefficient if the roads are congested, blocked, or otherwise dangerous because of the emergency. By not constraining evacuees to prespecified destinations, a one-destination evacuation approach provides flexibility in the optimization process. We present a framework for the simultaneous optimization of evacuation-traffic distribution and assignment. Based on the one-destination evacuation concept, we can obtain the optimal destination and route assignment by solving a one-destination traffic-assignment problem on a modified network representation. In a county-wide, large-scale evacuation case study, the one-destination model yields substantial improvement over the conventional approach, with the overall evacuation time reduced by more than 60 percent. More importantly, emergency planners can easily implement this framework by instructing evacuees to go to destinations that the one-destination optimization process selects.

Proposed Framework for Simultaneous Optimization of Evacuation Traffic Destination and Route Assignment

In the conventional evacuation planning process, evacuees are assigned to fixed destinations mainly on the basis of geographical proximity. However, the use of such prespecified destinations (an origin-destination table) almost always results in less-than-optimal evacuation efficiency because of uncertain road conditions, including traffic congestion, road blockage, and other hazards associated with the emergency. By relaxing the constraint of assigning evacuees to prespecified destinations, a one-destination evacuation (ODE) concept has the potential to improve evacuation efficiency greatly. To this end, a framework for the simultaneous optimization of evacuation traffic distribution and assignment is proposed. The ODE concept can be used to obtain an optimal destination and route assignment by solving a one-destination (1D) traffic assignment problem on a modified network representation. When tested for a countywide special event-based evacuation case study, the proposed 1D model presents substantial im...

AADT prediction using support vector regression with data-dependent parameters

Traffic volume is a fundamental variable in several transportation engineering applications. For instance, in transportation planning, the annual average daily traffic (AADT) is a primary element that has to be estimated for the year of horizon of the analysis. The huge amounts of money to be invested in designed transportation systems are strongly associated with the traffic volumes expected in the system, which means that it is important that the AADT should be accurately predicted. In this paper, a modified version of a pattern recognition technique known as support vector machine for regression (SVR) to forecast AADT is presented. The proposed methodology computes the SVR prediction parameters based on the distribution of the training data. Therefore, the proposed method is called SVR with data-dependent parameters (SVR-DP). Using 20 years of AADT for both rural and urban roads in 25 counties in the state of Tennessee, the performance of the SVR-DP was compared with those of Holt exponential smoothing (Holt-ES) and of ordinary least-square linear regression (OLS-regression). SVR-DP performed better than both methods; although the Holt-ES also presented good results.

Effects of Countdown Timers on Driver Behavior After the Yellow Onset at Chinese Intersections

Objectives: Few studies have focused on the effect of countdown timers at signalized intersections in China, where such timers are widely deployed for their perceived benefits of increased safety and capacity. This study examines the effect of countdown timers on driver behavior during the yellow interval. Method: Signal phasing and traffic operations were videotaped at 4 comparable signalized intersections under normal conditions. Microscopic details were extracted manually at 25 Hz to yield 24 h of data on onset time of the yellow, onset time of the red, driver location and actions after the onset of the yellow, red light–running violations, etc. For comparable intersections with and without countdown timers, driver behavior measured by driver decision (stop or go) and vehicle entry time (when the vehicle crosses the stop line) were analyzed using binary logistical regression (BLR) and a nonparametric test, respectively. Results: The results suggest that countdown timers can indeed influence driver behaviors, in terms of decisions to stop or cross the intersection as well as the distribution of vehicle entry times. There was a strong correlation between the presence of countdown timers and an increase in red light violations. Conclusion: Countdown timers may lead to increased entrance into the intersection during the later portions of the yellow and even the red. This alarming finding calls for further research as well as for serious consideration before the field deployment of countdown timers.

Improving Evacuation Planning with Sensible Measure of Effectiveness Choices: Case Study

This paper advocates the importance of selecting appropriate measures of effectiveness (MOE) for evacuation planning and optimization purposes. In addition to evacuation time, which is often the choice MOE used by many practitioners, this paper also presents several increasingly more complex and useful MOEs that consider average evacuee travel time, delay, and temporal–spatial-based exposure risks. By studying a real-world transportation network in the event of a hypothetical nuclear power plant incident, the paper demonstrates that the superiority of one evacuation plan (or an optimization strategy) over others is highly dependent on the MOE selected for evaluation. Optimization efforts based on a single and ill-chosen MOE may not yield the best evacuation plan, as one might expect. The overall space-based or time- and space-based risk exposure may be increased if the optimization process is focused on travel time or some other time-based measure alone. Depending on the specifics of an emergency scenario (e.g., disaster type, risk components, and time constraints), an appropriate MOE should be chosen judiciously to make sure the optimization process for the evacuation plan is not wasted. This paper also demonstrates a multiobjective optimization approach for planning and designing an evacuation plan. With combined space-based risk and the travel time in route searching and traffic assignment, better destination assignment (and improved evacuation efficiency) may be obtained. The findings here should provide insights for future efforts toward assessing, improving, and optimizing evacuation plans.

Estimating the Impact of Pickup- and Delivery-Related Illegal Parking Activities on Traffic

Illegal parking of delivery trucks used for pickup and delivery (PUD) reduces traffic capacity and causes delays. A geographically based combinatorial model was developed to estimate the extent of capacity losses and subsequent delays. The model uses a geographically based inference engine to extract data from several large-scale databases and process the data. These data are presented and compared with data from other temporary loss-of-capacity events. Because only weekday and daytime activities were studied, the resulting estimate of the national PUD effect is somewhat conservative.

Train–Vehicle Crash Risk Comparison Between Before and After Stop Signs Installed at Highway–Rail Grade Crossings

Objective: The safety benefit of stop sign treatment employed at passive highway–rail crossings has been a subject of research for many years. The objectives of this research is to investigate whether and to what degree the crash rate has changed at previously passive grade crossings after stop signs were implemented and examine whether and how the crash characteristics (associated with vehicle type, crossing surrounding, crossing design, crash severity, etc.) changed subsequently. Methods: Federal Railroad Administration grade crossing databases during the 26-year period (1980–2005) were applied in this study. Among the stop-controlled grade crossings, a total of 7394 “target” crossings were identified to be once crossbucks controlled and subsequently upgraded with the installation of stop signs without the implementation of other traffic control devices during the study period. Each target crossing was further divided into two time periods: when it was controlled by crossbucks only (before) and when it was controlled by stop signs (after). Both annual crash rate analysis and crash propensity analysis of before–after stop sign installation are conducted to quantify the safety benefit of stop sign treatment. Results: It was found that during the 26-year period (1980–2005), the annual crash rates when the crossings were controlled by crossbucks-only were consistently higher than the crash rates when the crossings were controlled by stop signs. The further crash propensity analysis indicated that the stop sign treatment was especially effective at crossings with higher annual average daily traffic (AADT), advanced warning signs, sight distance problem, adverse lighting conditions; the motorist-stopped-on-crossing, did-not-stop, and injury crash risks were also significantly reduced after stop signs were applied. Conclusions: The finding of this study suggested that the vehicle volume should be included into the guideline for stop sign use. Therefore, engineers and decision makers are encouraged to routinely check available sight distances at passive crossings controlled by crossbucks only and add stop signs to the crossings with insufficient sight distances. Additionally, it is suggested that advanced warning signs should be jointly used at stop-controlled crossings to maximize the safety effect. However, stop signs were less effective at crossings with higher train speeds or track classifications, where active warning devices may be a better safety solution for grade crossings.

Tracking Large Trucks in Real Time with License Plate Recognition and Text-Mining Techniques

Large-truck speeds on Interstate highways are not only a safety concern but have also become a problem with respect to air quality and energy consumption. Many cities have reduced truck speed limits to curb emissions. As potentially key building blocks of a proposed automated large-truck speed monitoring and enforcement system, license plate recognition (LPR) and plate-matching algorithms are the focus of this study. Present LPR systems are not perfect and can fail to read one-fifth to one-half of the characters on license plates, depending on various factors. Fortunately, even when LPR fails to read all characters on license plates, it is still able to read most of the characters with an appreciable degree of accuracy. By employing a text-mining technique called edit distance (with judicious use of travel time), this study demonstrates that it is possible to match 97% of license plates when only just more than 60% of the same plates were read correctly by LPR units at two different locations. This high matching rate does not entail a high false-positive matching rate either (about 2%). This paper presents the plate-matching algorithm in detail and provides statistics resulting from a field study conducted on I-40 in 2007. Several promising research directions for better matching efficiency and further reduction of the false-positive rate are identified.

Active and Passive Bus Priority Strategies in Mixed Traffic Arterials Controlled by SCOOT Adaptive Signal System: Assessment of Performance in Fortaleza, Brazil

In recent years, bus priority techniques for signals controlled by traffic management centers have become a viable alternative to reduce passenger delays at signalized intersections, especially in mixed traffic corridors. However, before any bus signal priority strategy is deployed in such corridors, the impacts on the different users of the system should be evaluated. The main objective of this work was to assess the operational performance of passive and active bus priority techniques in fixed and real-time signal systems of one of the main arterial corridors in Fortaleza, Brazil. As a secondary objective, it also evaluated the operational benefits of a SCOOT adaptive signal control system, comparing it with well-adjusted fixed-time plans optimized by TRANSYT, for periods of medium and high traffic volumes. In the evaluation of alternative scenarios, the following performance measures were considered: vehicle delay and number of stops simulated by SCOOT, as well as bus and automobile travel times observed in the field during the operation of each scenario. The results did not favor the adoption of passive and active priority schemes in the studied corridor; this led to the conclusion that SCOOTs real-time control, programmed for a good signal progression of the general traffic (buses and automobiles), is the best signal control strategy for an arterial corridor similar to the one under analysis.