Asphalt pavement distresses like surface shoving and slippage cracking can be found at airports in areas where aircraft brake and turn, such as high-speed exits, as a result of the high surface-shear forces. Slippage failure is typically caused by the deterioration of bonding between asphalt layers (delamination), or a lack of shear resistivity within the surface-layer asphalt mix. High pavement temperatures have also been shown to contribute to slippage failures in asphalt concrete pavements. At the intersection of Runway 4 R-22 L and High-Speed Taxiway N (HST-N) at Newark Liberty International Airport (EWR), interlayer delamination was determined to be the cause of shoving and slippage cracking on the pavement surface. In 2012, asphalt strain gauges were installed during a scheduled repaving of the runway and taxiway. This paper details the components of an asphalt strain gauge instrumentation system, and analyzes the strain responses collected from the gauges installed at EWR. By identifying large discrepancies in strain responses between strain gauges installed in the hot mix asphalt (HMA) overlay and lower layers of asphalt pavement (HMA milled surface), areas of potential delamination were identified. Delamination was successfully detected by the instrumentation and is shown to increase in severity over time, especially in the gauges nearest to the taxiway lead-line (centerline). Strain responses are also affected by temperature at the interface and aircraft speed. Photographs of the taxiway surface taken in the summer of 2014 confirm that slippage distress is occurring in the same areas in which the strain gauges indicate delamination. Statistical analysis methods were employed, and the Difference between Means tests and Kolmogorov-Smirnov Tests confirm that the measurement of $\mathrm{\Delta }$ strain ($\mathrm{\Delta }\epsilon$) was significantly distinct for bonded versus delaminated pavements.

In this study, a frequency-temperature correction model and an approach to construct the master curve of damaged asphalt concrete based on falling weight deflectometer back-calculated moduli were presented. Falling weight deflectometer (FWD) testing and field coring was conducted in 26 locations in Seoul City in which back-calculated and laboratory measured modulus were compared. A frequency-temperature correction model is developed based on the frequency-temperature superposition principle and was verified by comparing to current temperature correction models. The proposed model has the lowest prediction error of 8.2% in comparing asphalt concrete dynamic modulus, but is limited within the linear range of the dynamic modulus and reduced frequency curve in a logarithmic scale. A new procedure in establishing damaged asphalt concrete (AC) master curves is proposed considering a damage factor which is used in determining damaged $\delta$ and $\alpha$ parameters. It is verified using laboratory-measured modulus curves and is validated using FWD data obtained from literature. The validation study showed that the proposed procedures have an average error of 11.5%. Further study is recommended to enhance the proposed procedure to minimize error.

Developing crash-prediction models remains one of the primary approaches for studying traffic safety. Most of the current studies on single-vehicle (SV) and multivehicle (MV) crashes have only focused on the effects of exposure and geometric features of roadways and the effects of weather and traffic conditions are rarely incorporated. To provide more insightful observations, detailed weather and traffic data are adopted in this study. As a result of adopting detailed data, multiple daily observations are generated for SV and MV crashes on each roadway segment, forming a multivariate panel data set that poses some methodological challenges. A new approach to analyze SV and MV crashes is proposed by developing a bivariate Poisson lognormal model with correlated segment-specific random effects. The proposed model can characterize both the multivariate and panel nature of the data, and readily address the following three types of serial correlations within the multivariate panel data used in this study: (1) correlation between SV and MV; (2) temporal correlations across time within each segment for SV and MV, respectively; and (3) possible connection between temporal correlations for SV and MV crashes. It is found that the proposed model outperforms the two competing models by addressing as much unobserved heterogeneity as possible, dealing with excessive zeros in the observed data, and bearing the smallest deviance information criteria value. The proposed methodology is applied to a mountainous freeway section on I-70 in Colorado, where the climate is subjected to rapid change due to high elevation. Results show that weather- and traffic-related explanatory variables, especially surface conditions, play a significant role in affecting the occurrence of SV and MV crashes. Moreover, differences between contributing factors for SV and MV crashes are also investigated.

In this paper, for a quarter-car model, root-mean square response variables are calculated for roads classified by the international standard organization, based on the straight-line spectrum models. Covariance expressions are derived for vehicle response variables assuming that the road spectrum is that of a Motor Industry Research Association (MIRA) spectrum and compared with the existing results in the literature. The international standard and the MIRA spectra are then compared with the models delivered by high resolution frequency-domain subspace identification algorithms. The IRI statistic of the quarter-car response is calculated for the road models built by the curve fitting and the identification methods from the measured road spectrum assuming that the road excitation is a zero-mean Gaussian process

Vehicle following has always been a hotspot of research in the transportation system. It can be used in many applications, such as unmanned ground vehicle (UGV) queue marching, traffic congestion relief, automatic driving systems, etc. Trajectory tracking has been the main consideration in previous studies, but the research about following distance is rare. In this paper, a novel vehicle-following model is proposed for independent vehicle following. In the proposed model, multisensor information is fused to promote the detection accuracy. Moreover, a virtual piecewise spline tow-bar is built between the leader vehicle and the follower vehicle. According to the state of virtual tow-bar, the following speed and heading can be well controlled, to keep a reasonable gap between two vehicles. The proposed vehicle-following model is designed for low speed and minor vehicle spacing case. By using the proposed vehicle-following model, the performance of multple-vehicle automatic cruise can be improved, resulting in a steady traffic flow on an express way. Then the stop-and-go waves of congestion can probably be relieved. Simulation and experiment are conducted to validate the efficiency and robustness of the proposed vehicle-following model.

As the demand for additional capacity on urban freeways continues to increase, there is a need to better understand how to optimize the lane and shoulder configurations so as to improve capacity without adversely impacting operating speed. The overall objective for this research effort was to identify the operational implications of using reduced lane and shoulder widths on freeways. The operational analysis found speeds $3.5\mathrm{km}/\mathrm{h}$ (2.2 mph) higher for a 3.7 m (12 ft) lane as compared to a 3.4 m (11 ft) lane. The authors developed an equation that can be used to predict speeds per lane for freeway locations with boundary conditions similar to those present in the study database.

A new method for track inspection is developed to complement traditional geometric methods. An inertial measuring system and a specialized data treatment method is proposed to evaluate railway track safety, observed from the vehicle dynamic performance point of view. System equations for the inverse vehicle dynamic problem, augmented with the suspension torsion equation, are solved to estimate the wheels driving forces that are directly correlated with the vehicle safety traveling over an irregular track. Results of a measuring campaign on a railway wagon are used to evaluate the present system by direct positional comparison with lateral and vertical contact forces ($L/V$) safety ratio measured with two instrumented wheelsets (IWS) installed on the leading bogie of the wagon. Also, data collected and treated are compared with track geometry measurements. The results of quantified track safety present good agreement with those found using traditional measuring methods. This confirms the ability of the new method to detect the location of the highest potential hazard region, for optimized track maintenance purposes. The new system proves to be a promising method for track safety evaluation.

Maintenance, repair, and rehabilitation (MR&R) are cost-effective activities for maintaining good serviceability of concrete bridge decks during their lifetimes. Many past studies were conducted to provide public agencies with optimal MR&R strategies for selecting the correct treatment types at the appropriate application times. While researchers did establish and confirm that the annual deterioration rate and the effectiveness of the MR&R activities are the key input factors that affect the outcomes of MR&R strategy models, they did not spend nearly as much time on engineering inquiry to measure the effectiveness of MR&R activities as they did on projecting the deterioration patterns of bridge systems. The issues of selection bias, empirical data reliability, and overstatement of the effectiveness of MR&R activities continue to compromise the present methods. Therefore, this paper presents a new method that addresses these issues by utilizing the bridge condition evaluation criteria commonly applied among state DOTs in the United States. The new method’s model was tested using a case study that verified its usefulness for a state DOT to establish reliable, practical MR&R strategies.

The roadside of horizontal curves should be cleared of any obstructions to satisfy sight distance (SD) requirements. This paper proposes an innovative roadside design curve for the required lateral clearance (offset) on simple horizontal curves. The roadside curve consists of a circular curve connected with two spiral curves and exhibits several advantages: (1) it provides a smooth SD profile that improves design consistency; (2) it has beneficial effects on safety for sharp horizontal curves; and (3) it leads to the development of a simple design aid. The roadside curve provides a good approximation of the required lateral offset for roadway simple horizontal curves with a length lesser or greater than the required SD. The results show that approximately two-thirds of the maximum lateral offset required by the American Association of State Highway and Transportation Officials should be provided at half of the required SD upstream and downstream of simple horizontal curves. The concept of roadside curve is similar to the design of spiraled horizontal curves and thus will be easily understood by highway designers.

Differential car/truck speed limits have been used to enhance road safety and increase the fuel efficiency of trucks. For two-lane highways, the safety of these differential speed controls could be affected by possible changes in the pattern of car and truck interactions and their associated crash risks, especially as it pertains to overtaking. In this paper, a microscopic simulation model is presented that assesses the safety and traffic implications of differential car/truck speed limits for two-lane highway operations, with emphasis on the overtaking maneuver. Three speed control strategies are considered: uniform posted speed limits (USLs), differential posted car–truck speed limits (DSLs), and differential mandated truck speed limits (MSLs). Mandated limits involve the use of on-board truck limiters or governors with preset maximum thresholds. The results from several simulation tests suggest that DSLs and MSLs reduce the average travel speed (ATS) of the traffic stream. DSL and MSL controls were found to slightly increase head-on time-to-collision (TTC) and percentage time spent following (PTSF). The total number of overtakes was found to increase slightly for DSL and MSL controls compared to USLs. While the number of car–truck overtakes increased significantly for DSLs and MSLs, the number of car–car overtakes resulted in a significant decrease. It was concluded that three measures of ATS, TTC, and car–car overtaking positively impacted safety, while PTSF, car–truck, and total number of overtakes negatively impacted safety because of differential speed limit controls.

In transportation studies, modeling human learning and decision-making processes plays a key role in developing realistic safety countermeasures and appropriate crash-mitigation strategies. In this study, a human learning model was created that captures the cognitive structure of human memory. The relationship between long-term and short-term memories was incorporated into a reinforcement learning technique to construct the human learning model. The model was then applied to dilemma zone data collected in a simulator study. Dilemma zone is an area of roadway ahead of the signalized intersection in which drivers have difficulty deciding whether to stop or proceed through at the onset of yellow. Driver choice behavior and learning process in dilemma zones was modeled, taking into account drivers’ experiences at the previous intersections, and was compared to a pure machine learning model. The results of the model revealed lower and faster-merging errors when human learning was considered in training agents. The human learning model for dilemma zones presented here could be used to evaluate dilemma zone mitigation algorithms by considering their effects on driver agents.

Nonstructural factors, such as surface distresses and ride quality, have been commonly used as the main indicators of in-service pavement conditions. In the last decade, the concept of implementing a structural condition index in pavement management system (PMS) to complement functional condition indices has become an important goal for many highway agencies. The rolling wheel deflectometer (RWD) provides the ability to measure pavement deflection while operating at the posted speed limits causing no user delays. The objective of this study was to develop a model to predict pavement structural capacity at a length interval of 0.16 km (0.1 mi) based on RWD measurements and to assess its effectiveness in identifying structurally deficient pavement sections. Rolling wheel deflectometer data collected from 153 road sections (more than 1,600 km) in District 05 of Louisiana were used in this study. The predicted structural number (${\mathrm{SN}}_{\mathrm{RWD}0.16}$) showed an acceptable accuracy with a root-mean square error (RMSE) of 0.8 and coefficient of determination ($R^2$) of 0.80 in the validation stage. Core samples showed that sections that were predicted to be structurally deficient suffered from asphalt stripping and material deterioration distresses. Results support that the developed model is a valuable tool that could be used in PMS at the network level to predict pavement structural condition with an acceptable level of accuracy.

Past studies have determined the effects of the pavement mechanistic-empirical (ME) default (Level 3) values of vehicle class distribution (VCD) and axle load spectra (ALS) on pavement performance. However, it is still not clear how the clustered VCD and ALS affect the ME predicted pavement performance. In this study, traffic data from 10 weigh-in-motion (WIM) stations were gathered and analyzed to develop the VCD and ALS values using arithmetic average and clustering methods (Level 2). Next, using Level 2, Level 3, and site-specific (Level 1) inputs of VCD and ALS, the pavement ME predicted performance was determined. The results show that the predicted performance by the cluster (Level 2) data are very close to those of the site-specific data (Level 1). Performance generated by the ME default values (Level 3) are significantly different from those generated by the site-specific or cluster values. When comparing the performance of the ME design default (Level 3) with those of the statewide average data, the ME design default VCD produces fewer errors than the ALS. This study recommends using clustered or site-specific WIM data instead of ME default or statewide average value.