Highway designers occasionally use unsymmetrical vertical curves in locations where the vertical clearance controls. In this case, the symmetrical curve, which has a constant rate of change of grade, is replaced with two arcs, which have two different rates of change of grade and a common point of tangency. Traditional unsymmetrical curves are designed so that the point of common tangency (point of compound curve, PCC) is at the same station of the point of intersection between the curve's two tangents. However, a better unsymmetrical curve can be achieved by locating the PCC in the middle of the curve. The resulting curve, called the equal-arc unsymmetrical (EAU) curve, enhances sight distance, highway aesthetics, driver comfort, and vertical clearance. In this paper, the sight distance profile on the EAU curve is analyzed and compared to that of the traditional curve. The EAU curve was found to improve the minimum available sight distance compared to the traditional curve, and thus the required length of an EAU curve to satisfy a specific sight distance would be shorter than that of a traditional curve. Also, design length requirements for the EAU curve that satisfy stopping (daytime and nighttime), passing, and decision sight distances are developed.

Growing awareness exists of the important role of consistency in highway alignment design in minimizing the frequency and extent of violations of driver expectancy, to improve highway safety. A simple and practical safety margin, based on an established driver behavior model, is used to investigate two sources of inconsistency: (1) The choice of design speed; and (2) the choice of superelevation and side friction when the curve radius exceeds the minimum. This paper describes a method for choosing the design speed for a horizontal curve, taking into account the extent of the curve. The paper also describes alternative methods for choosing the superelevation and side friction. Using a practical numerical example, the design speed selection and superelevation and side friction selection methods are evaluated with respect to consistency of the safety margin. It is concluded that use of the design speed selection method, and having the superelevation and friction varying linearly with the degree of curvature, will lead to greater alignment consistency and improved safety.

A time based microsimulation model is developed for analyzing the traffic operation at single point urban diamond interchanges. Features of the model include actuated signal operation, protected and permitted left turn phasing, right turn phasing with and without right turn on red, traffic in shared lanes, traffic in left turn and right turn storage lanes, car following, lane changing, gap acceptance behavior, primary and secondary queue formation and dissipation. The model accepts geometric, traffic, and signal data in an interactive mode. Input files may also be created separately without going through the interactive session. The model outputs include measures of effectiveness such as stopped delay, total delay, average speed, and maximum and average queue length. These measures of effectiveness are given for each turning movement, for each approach, and for the intersection as a whole. The model outputs also show the total green time and the total yellow and all red times assigned by the actuated system to each phase during the simulation time. The model results are compared with the actual data collected in the field.

Analytical models are developed to optimize bus service with their demand characteristics. The optimizing bus services design options are analyzed in this paper, while route spacing and headway are included with or without vehicle capacity constraints. Optimal levels of these variables are obtained analytically for the objective functions of profit maximization. The analysis uses an equilibrium framework with transient ridership sensitive to the level of service provided by the bus system. The major results consist of closed-form solutions for the optimal area operating policies that are dependent upon local conditions (vehicle capacity constraints) and coefficients of the demand.

Airfield pavement evaluations commonly include nondestructive testing. These evaluations are used for projecting pavement performance and for designing pavement rehabilitation options. To apply reliability concepts to these prediction and design scenarios, it is necessary to learn more about the variabilities associated with nondestructive testing and pavement evaluation procedures. In this study, twelve sites around the southeastern United States were tested with both a falling weight deflectometer and a heavy-weight deflectometer. The data were evaluated using the military procedure. The test data and the pavement evaluation results were examined for spatial variability and for test device variability. Two primary conclusions resulted from this study. The first primary conclusion is that the current military evaluation procedure does not provide a consistent level of reliability for evaluation results. The second primary conclusion is that the magnitude of the load used for deflectometer testing affects the evaluation results.

This study presents a method of estimating the modulus of elasticity of an existing multilayer pavement system. The method is based on the interpretation of a measured deflection basin. The deflections of the existing pavement were measured using the tiltmeter, a new device of high resolution. The measured deflections are to be matched with deflections from a computer output that corresponds to a specific combination of thicknesses and normalized moduli. The computer program presumes that each layer is linearly elastic, and that the moduli be constant at all points in the layer. The method provides an easy and precise way of estimating the pavement moduli.

A new elastoplastic model is presented for predicting the stress/strain response of asphalt concrete under cyclic loading. The model utilizes multiyield surfaces and isotropic hardening. Rowe's stress dilatancy theory is used to obtain the relationship between the permanent volumetric and vertical strains as well as the hardening law for the changes in the sizes of the plastic moduli caused by cyclic loading. These relationships considerably simplify the task of predicting the plastic deformation under cyclic loading of both triaxial compression and extension tests. A complete description of the plastic deformation of asphalt concrete under cyclic loading is treated, including the elastic, viscoelastic, and plastic components as well as the relationship between rutting and cracking. A computer program, called Rutting in Asphalt Concrete, RUT, determines the plastic deformation under cyclic loading. There is good agreement between the model predictions and experimental data that includes data published by other researchers.

For different overlay systems currently used in practice (PGM 14, PGM-G, and TENSAR grid) the fracture behavior of the interface between interlayer and asphalt overlay as well as of the entire system is characterized and analyzed. A simple and easy-to-use wedge splitting test according to Tschegg is carried out in order to evaluate both the interface and reflective cracking behavior. The specimens used in this test consist of modified cores taken from road pavements. This investigation demonstrates that the overlay systems show considerable variation in their fracture mechanism parameters (such as resistance to crack propagation of the interface or propagation of reflection cracks), but only minor differences with regard to the other properties (such as bonding tensile strength of the interlayer compound system or the notch tensile strength of the reflective crack). The effectiveness of the interlayer system is diminished when the resistance against crack propagation in the interface is reduced.

Laboratory investigations were carried out on a residual lateritic soil treated with quicklime (up to 8% by weight of dry soil) in order to evaluate the effect of lime content, curing period, and compactive effort on the permeability of lateritic soil-lime mixtures prepared at various maximum dry densities and corresponding optimum moisture contents. The permeability of uncured specimens (standard Proctor) increased to a maximum at 4% lime content and decreased with increasing lime content. Specimens compacted at the energy of the West African Standard had coefficients of permeability that decreased with increasing lime content. For the cured condition, the permeability of all the lime-treated specimens compacted at the two energy levels increased with curing age up to 14 days and decreased with curing age beyond 14 days.

Dynamic loads caused by vehicle-pavement interaction are either moving loads or random loads. Motion property of dynamic loads was described by means of Dirac-delta and Heaviside functions, and its randomness property was studied using a quarter-truck vehicle model. A series of statistical characteristics of wheel loads in frequency domain and time domain are presented in this paper. It was found that dynamic loads are a constant mean Gaussian stationary ergodic process whose power spectral density is in proportion to the power spectral density of pavement roughness, whereas the ratio coefficient relates to dynamic properties of vehicle and velocities of travel. Numerical simulation based on the specifications of the International Organization for Standardization (ISO) indicated that only at low speed could dynamic loads be treated as quasi-narrow band-width random processes, which covers the frequency range of 1.0–4.0 Hz. At high speed, main frequency distributions of dynamic loads occur in 1.0–4.0 and 10.0–15.0 Hz. Further analysis also showed that according to the International Organization for Standardization specifications, the root mean square of dynamic loads is proportional to $C_{\mathrm{SP}}\text{,}$ the square root of an important index related to pavement levels.

This paper presents the test results from full-scale accelerated pavement testing with the Texas Mobile Load Simulator. Data from in-situ instrumentation and nondestructive testing were collected and analyzed at different loading stages to assess material property changes under accelerated loading. Forensic studies were performed at the end of trafficking to assist in explaining the performance of the pavement. Two trenches were made to study material characteristics in the longitudinal and transverse directions. It was found that at the early stage of trafficking the test pad responded to falling weight deflectometer (FWD) load linearly, not only over the whole pavement system but also within individual layers. Before mobile load simulator testing, FWD data indicated the weakest area exists at the left wheel path (LWP) of 7.5-m line (7.5L). Later, this weak area was confirmed to have the highest rutting and the most intensive cracking. The dynamic cone penetration results showed that the base at this location was at its weakest. Also, at 7.5L the dry density was lowest, ∼7% lower with a moisture content ∼8% higher than the adjacent area. The LWP had higher FWD deflections than the right wheel path (RWP), and consequently the LWP manifested more rutting. This proved to be primarily due to differences in moisture content. This was probably because more water infiltrated in this area during rain due to manifestation of more extensive cracking during early phases of trafficking. The maximum surface deflection values increased as trafficking increased in the left and right wheel paths due to pavement deterioration, while deflection for the center remained constant because of the lack of traffic loading. The LWP had more rutting than the RWP and this correlated with the measured FWD deflections prior to trafficking. The W1 values increased as trafficking increased for the LWP and RWP due to pavement deterioration. The majority (>60%) of rutting was from the 300-mm uncrushed river gravel base.

This study illustrates the usefulness of the finite-element method in the analysis of three-layer pavement systems subjected to different types of loading. The method is capable of simulating the observed responses of pavements subjected to axle loads with different tire pressures, axle loads with different configurations, and axle loads traveling at different speeds. A variety of material constitutive models such as linear elastic, nonlinear elastic, and viscoelastic are employed in the analyses to describe the behavior of the pavement materials. Finite-element modeling of pavements, if validated, can be extremely useful, because it can be used directly to estimate primary response parameters without resorting to potentially costly field experiments. If accurate correlations between the calculated and the measured primary response parameters can be obtained, then the analytical model can be used to calculate primary response load equivalency factors, utilizing deflection-based or strain-based equivalency factor methods.

The effectiveness of guardrail installations and other road modifications to reduce injuries has been questioned recently based on critiques of study designs comparing injuries before and after installation at high risk sites. So-called “regression to the mean” and “accident migration” are claimed to greatly reduce previous effectiveness estimates. In this study, fatalities on 161 km of road from the 7 years before and 10 years after guardrail installation at selected sites were compared, noting where installations occurred and how many people died when vehicles went over embankments. Regression analysis was used to estimate the effects of the guardrails, seat belt law, average daily traffic, and trend. In the 10 years after installation, 21 fatalities were expected but no fatalities occurred at new guardrail installations, an average reduction of about two per year. Other factors were not significant. Based on fatalities on the interspersed sections of road with no new guardrail, there is no evidence of “regression to the mean” or “accident migration.”