This paper presents an analytical model for prediction of shrinkage-induced stresses and displacements in concrete pavements due to the restraint of the supporting base. The model relies upon the interfacial friction stress transmitted between the slab and base as the fundamental mechanism of restraint provided by the base when concrete shrinks. By analyzing experimental results reported in the literature, a bilinear relation describing the friction stress-slippage behavior of the slab/base interface is adopted in the model. Closed-form expressions of shrinkage stresses and displacements in the concrete slab are obtained. The model quantifies the effects of concrete properties and the interfacial characteristics of the concrete slab and base on the shrinkage stresses and displacements. By applying the slab/base frictional stress-slippage relation of seven typical supporting bases in the model, the influences of the supporting base on the development of shrinkage stresses and displacements in the concrete slab are investigated. In addition, the effect of “bond breaker” layers on the shrinkage stresses in the concrete slab lying on a hot mix asphalt concrete base is studied with the model.

This paper presents a theoretical model for determining international roughness index (IRI) and present serviceability index (PSI) in term of power spectral density (PSD) roughness. Frequency response functions of quarter-car model and CHOLE-type profilometer are obtained by means of stochastic process theory. Correlation of IRI and PSD becomes possible based on their PSD-based expression. It is also shown that statistical regression may not be effective when attempting to correlate IRI with PSI due to their inherent multivalued corresponding property. Transportation agencies can use PSD-based models to precisely convert IRI to PSI given that PSD roughness of a pavement is known. Furthermore, this paper provides a correlation between typical PSI and IRI of flexible and rigid pavements whose PSD roughness is given in the form of power law at specified levels.

Because of the lack of formulas for evaluating the peak value of dynamic vehicle load on rigid pavement, the need to derive a simple closed-form solution to predict the peak load for pavement design is apparent. Presented herein is a theoretical solution for evaluating the probable peak vehicle load on rigid pavement due to the passage of a vehicle moving at constant speed along a rough road surface. The analytical approach is based on the application of a road displacement spectrum imposed at the lower end of a vehicle system to model the road surface roughness. The engine-induced force is also considered in analysis. Based on the proposed solutions, relations of design vehicle load and traffic volume for discrete risks are constructed for applications of pavement design.

The objective of this study is to experimentally and theoretically estimate the strain distribution in an asphalt binder as it exists in hot mix asphalt (HMA). This is important to evaluate whether the strain magnitudes can cause nonlinear behavior for the binder and HMA. The experimental procedure relies on capturing images of the surface of an HMA specimen during deformation. A computer algorithm is then used to calculate the strain values within an image. Finite-element analysis of the internal structure of HMA is also used to estimate the strain distribution. The results are shown to have good correlation with the experimental measurements. However, due to limitations imposed by the image resolution, the strain distribution is computed within areas that include a combination of binder and fine particles (mastic) rather than the binder phase. Consequently, micromechanics analysis of the mastic is used to calculate the binder strain. The results demonstrate that the binder strain magnitudes could reach high values well into the range of nonlinear behavior of the binder.

This paper reports the findings of a project on evaluating the performance of recycled concrete aggregate (RCA) for use as a base material under hot mix asphalt pavements and as an aggregate in portland cement concrete pavements. Nine design sections involving hot mix asphalt and portland cement concrete pavements were constructed at the University of Central Florida's Circular Accelerated Test Track to evaluate the performance of the pavement sections made with RCA under actual dual-wheel loading. Laboratory tests performed on concrete specimens made of different percentages of RCA and virgin aggregate (VA) showed that the higher the percentage of RCA material, the lower the compressive, flexural, and tensile strengths. Despite having lower compressive strength than VA concrete, 100% RCA concrete had a compressive strength of 35 MPa (5,070 psi), which was well above the target strength of 25 MPa (3,630 psi). Trenching of the flexible pavement test sections and petrographic examination of the hardened concrete showed that after a simulated life expectancy of 3.04 years for asphalt sections and 4.2 years for concrete sections, RCA has performed at least equally to VA.

A macroscopic pavement management system has been designed to yield optimum pavement conditions under constrained budgets. The system can predict future pavement conditions from performance curves generated using the serviceability concept developed by the American Association of State Highways and Transportation Officials (AASHTO). The pavement performance curves are generated using an incremental analysis of the AASHTO basic design equation. The system then applies an effective optimum decision policy that aims toward maximizing the annual average network present serviceability index according to the expected outcomes of predefined major rehabilitation actions and subject to budget constraints. In contrast to the microscopic approach that considers “short” pavement sections, the macroscopic approach to pavement management uses “long” pavement sections, i.e., projects. Each project is constructed of the same pavement structural section and possesses a unique performance curve. Pavement projects with similar performance trends are grouped into six classes with each class treated as a single entity. The system is particularly beneficial in establishing long-term pavement rehabilitation policies in terms of planning, scheduling, and budgeting.

To maximize the potential for success of an urban high-occupancy vehicle (HOV) lane, there are several design intricacies that require special attention, particularly at critical intersections. This paper is based on extensive research of HOV lanes, specifically their optimal setback planning and the resulting impact on delay, that was conducted recently in Israel. A setback, the termination of an HOV lane prior to a critical intersection—typically characterized by long delays and diverging queues—allows passenger vehicles, in addition to buses, to use the HOV lane. Throughput capacity is thus increased and delay for passenger vehicles reduced at the critical intersection, especially when the HOV lane is not utilized to its fullest capacity. The paper also presents the geometry of a setback, with and without a right-turn lane and, in addition, introduces the development of a model for determining the optimal length of a setback. Average delay to HOV vehicles and to non-HOV vehicles is given as a function of the degree of saturation for a design that both includes and does not include a setback. At critical intersections, the provision of an HOV lane with an optimal setback is shown to result in the least amount of total passenger delay.

This paper presents a calibration-free model to estimate the average speed of moving vehicles in a link, from the upstream intersection until 40–60 m from the downstream intersection's stop line. The proposed model calculates link speed from travel time, which in turn has two components: cruise time and signal delay. The cruise time is based on the speed measured by loop detectors. The signal delay is first calculated using the Webster formula. A coefficient is derived to portion the signal delay to the link segment upstream from the stop line. This model, which takes input from loop detectors and signal timing parameters, has been tested against simulated data. The results show that the model is able to provide link speed estimate to within 5 km/h error at least 95% of the time. This accuracy is better than that of the British and Iowa models and is only slightly lower than the Illinois model. However, being calibration-free, the proposed model here has a distinguished advantage in practical implementation.

Black spots (BS) are highway locations where the potential for accidents is unacceptably high when compared to the established risk tolerance criteria. It is argued in this paper that to effectively utilize available funds for safety improvements, one must first designate high priority BS locations. This paper presents the results of two model applications for establishing the potential for accidents and designating safety BS along a highway. The two models are based on multivariate Poisson regression and empirical Bayesian (EB) methods. The application involves a 25-km stretch of highway in southern Italy, for which accident and exposure data are available for the period 1993–1999. The pattern of BS suggested by each model is compared. The EB model was found to yield fewer BS locations than the Poisson regression model. If, as argued in this paper, safety countermeasures are best applied at BS, use of the EB model could result in significant cost savings.

During the period of September 1999 through March 2000, the University of Louisville Research Foundation conducted a series of reference energy mean emission level (REMEL) measurements of individual vehicle operations in the state of Arizona. This study was conducted to determine if the noise emission characteristics of Arizona highway vehicles are consistent with the data contained in the Federal Highway Administration's Traffic Noise Model (TNM). A comparison of the Arizona TNM REMELs with the Federal Highway Administration national TNM REMELs shows the two data sets are statistically different at a 95% confidence level for all three vehicle types addressed in this paper.

There is a requirement in real-time, routing information systems to identify the “optimal” route based on the multiple objectives and the individual decision-making rules of the users. While a number of utility theory-based techniques have been developed to accomplish this goal, there are a number of shortcomings regarding implementation. These problems, in particular, are: (1) obtaining a mathematical representation of the driver's utility function for each trip type is difficult in practice; (2) the drivers' utility function is a function of the choice set and is difficult to calibrate prior to the identification of the choice set (i.e., it is context-dependent); and (3) identifying the optimal path using a realistic multiple attribute nonlinear utility function is a nondeterministic polynomial time hard problem. Consequently, a heuristic two-stage strategy that identifies multiple reasonable routes and then selects the “near-optimal” path may be an effective and practical alternative. The second step of this proposed strategy is the focus of this paper. A fuzzy logic-based multiple objective route choice model (or decision support system) is developed in order to evaluate the alternative routes identified in the first step. Fuzzy logic is used to take into account crisp values, fuzzy numbers, and linguistic variables that are common phenomena in a real-time vehicle routing environment. The routes and the route attributes are identified and a posterior utility function is developed by combining the prior utility function with a context-dependent utility function that is derived from an entropy model. The proposed strategy is illustrated using Intelligent Transportation System (ITS) data from Houston, Texas, and tested on a traffic network from Austin, Texas, under various traffic conditions. When multiple attributes were considered, an alternative path to the fastest path was found to be the “optimal” path for a significant number of O-D pairs, and this was likely to occur as the level of congestion and the O-D distance increased.

Border crossings in northern New York are important to international trade and tourism. How traffic operations at these crossings can be made more efficient has been a concern to the New York State Department of Transportation and the agencies that operate bridges, toll plazas, and customs and immigration plazas. Very little, however, is known about the characteristics of vehicle-processing time and delay at these crossings. This lack of understanding makes it difficult to evaluate the potential impact of changes in operating policies and physical plants. To alleviate this problem, this study identifies the processing time characteristics at three northern New York border crossings and proposes a delay model for planning applications.