Critical length of grade control, fixed-elevation points, and nonoverlapping of horizontal and vertical curves are three common requirements in the vertical alignment design of roads. These three forms of constraints are, however, usually not addressed in the conventional road alignment optimization analysis because of the complexity in considering them in the mathematical formulation and solution of the problem. This paper illustrates that the artificial intelligence technique of genetic algorithms can be adopted to handle these three forms of constraints effectively. The formulation of the genetic-algorithm computer program and the method of solution are explained. The validity of the optimization algorithm is verified against a dynamic programming solution. Examples are presented to illustrate the application of the genetic-algorithm program to problems involving critical length of grade requirements, fixed-elevation control, and nonoverlapping of horizontal and vertical curves. These three constraints were found to have significant effects on the computed optimal alignments and the associated construction costs.

Estimation of parameters in the distribution of headways in roundabouts is a cumbersome task. Inherent to data employed are temporal variation together with dependence between successive vehicles. This is a consequence of the sampling period and interactions in the roundabouts. The elimination of observations, so as to obtain trend-free samples of independent vehicles, removes too much information from the data. If all the observations are used, the results of the parameter estimations are affected by temporal variation and dependence between headways. In this paper, an analysis of the statistical properties of the headways between major stream vehicles in roundabouts is described. The properties investigated were the existence of temporal variation in the measured headways and independent and identically distributed successive headways, i.e., if the headways were from a random sample. More than 50% of the investigated subsets suffered from trends or other nonrandom variations. The question whether these subsets should be excluded or not is discussed. Three methods for the estimation of parameters in the M3 distribution were tested. For the two-lane subsets, it was possible to demonstrate a linear relationship between the proportion of free vehicles and the volume. For the one-lane subsets, the adjusted $R^2$ was low, and the linear constant was not significant. Nevertheless, the distributions estimated by use of the two-lane subsets did not provide a better fit to the observed data than the distributions estimated by the use of one-lane subsets. The predicted capacity for a minor lane, i.e., the capacity obtained by the fitted distribution, was close to the capacity obtained by the empirical distribution function (the estimated capacity).

This paper introduces the concept of a linearly increasing variable power to the basic vehicle dynamics model that has been proposed in the literature. The paper also calibrates the model to typical truck classifications 8–10 that travel along interstate highways in the United States. The proposed enhancement is demonstrated to result in significant improvements in vehicle performance curves, especially at low speeds when vehicles are engaged in gearshifts.

In congestion pricing, apart from the conventional social equity issue between poor and rich drivers who pay the same toll charge, there exists a spatial equity issue in the sense that the changes of the generalized travel costs of drivers travelling between different origin-destination (O-D) pairs may be significantly different when tolls are charged at some selected links. The former has been debated extensively, whereas the later is blatantly ignored in the literature. In this paper, we propose bilevel programming models for the network toll design problem by explicitly incorporating the social and spatial equity constraints in terms of the maximum relative increase of the generalized equilibrium O-D travel costs between all O-D pairs for various classes of drivers with different values of time. A penalty function approach, using a simulated annealing method, is applied for solving the equity-constrained toll design problem and demonstrated with a simple network example.

Transit operations are interrupted frequently by stochastic variations in traffic and ridership conditions that deteriorate schedule or headway adherence and thus lengthen passenger wait times. Providing passengers with accurate vehicle arrival information through advanced traveler information systems is vital to reducing wait time. Two artificial neural networks (ANNs), trained by link-based and stop-based data, are applied to predict transit arrival times. To improve prediction accuracy, both are integrated with an adaptive algorithm to adapt to the prediction error in real time. The bus arrival times predicted by the ANNs are assessed with the microscopic simulation model CORSIM, which has been calibrated and validated with real-world data collected from route number 39 of the New Jersey Transit Corporation. Results show that the enhanced ANNs outperform the ones without integration of the adaptive algorithm.

The stopped delay is a primary factor for measuring the level of service at signalized intersections, and the 1994 Highway Capacity Manual proposed a method for measuring the stopped delay in the field based on direct observation of “stopped-vehicle counts.” This paper evaluates the expected error in delay measured by this method for various ranges of counting intervals and survey offsets. A mathematical formulation was derived to evaluate this error for different combinations of relevant parameters. Results indicate that the expected error may exceed ±100% of the deterministic delay, and significant errors are generally associated with shorter cycles, higher green to cycle ratios, lower degrees of saturation, and longer intervals. Significant errors are also reported when the interval perfectly divides the cycle length. A set of optimum offsets is determined for use with the 20-s interval and different combinations of other parameters. These offsets confine the expected error in the measured stopped delay to be within ±1%, a significant improvement regardless of whether the cycle length is divisible by the interval.

This paper demonstrates the feasibility of an advanced concept of traffic control for congested urban arterials in an intelligent transportation system (ITS) setting. The core idea of the new concept is to integrate the capability of dynamic adaptive signals with dynamically optimized time-dependent variable speed. Under such control, speeds would be automatically optimized and set by a central computer in parallel with the other signal control parameters. Speed and signal control parameters would change between links and over time, in response to changing traffic conditions. Drivers would follow the optimized speed as they enter a link. Once an optimal speed has been set for a link, it remains constant until the control cycle ends. Link speeds would be updated only at the end of every control cycle. The control cycle may change in length as system conditions evolve. The new control concept was tested on a congested arterial with multiple links. The arterial system was modeled as a discrete event time varying dynamic system with a control period spanning several cycles. System throughput was maximized subject to such critical operational measures as intersection blockage, queue spillbacks, and other relevant traffic operation measures. Genetic algorithms were used as an optimization tool. Results show that the proposed control concept will significantly improve traffic flow. The new control concept is suitable for on-line implementation in an ITS setting.

The yellow change and red clearance intervals for left-turn movements in right-hand-side driving environments are addressed on an analytic basis. Factors such as safety, perception, comfort, driver’s behavior, signal timing, and traffic ordinances are taken into account for setting the intervals. The yellow change interval for left-turn movements is found to be in general longer than that required for straight movements at intersections, because motorists usually slow down before making turns. Moreover, the required red clearance interval is derived in terms of the average curvature of a left-turn curve and the magnitude of acceleration that motorists are willing to bear on the curve. The calculated yellow change and red clearance intervals are in good agreement with the field observations for two signal-controlled intersections. The red clearance intervals in practice are shown to be short in comparison with the field observations and analytic estimates.

To increase the capacity of through traffic at signalized intersections, additional lanes with limited lengths—called auxiliary lanes—are usually added to the roadway at the intersection. Because of their limited length, as well as other factors, these lanes are not fully utilized as other continuous through lanes. The objectives of this research were (1) to identify the level of utilization of auxiliary through lanes added at intersections of four-lane roadways with right-turn driveways and/or side streets present within the downstream length of the auxiliary lane; and (2) to study the effects of auxiliary lane length, right-turn volume exiting downstream into driveways/side streets, and through/right-turn lane group delay on the level of their utilization. Lane-use and right-turning volume data collected during 866 saturated cycles at eight signalized intersections with different auxiliary lane lengths and right-turn volumes were used to accomplish the research objectives. Auxiliary lane length and right-turn volume downstream were found to significantly contribute to the use of auxiliary lanes at the 0.01 level of significance. Lane-use of nearly 0.4–1.9 straight-through vehicles per cycle, depending on lane length and downstream right-turn volume, was observed at the study locations. Though not directly comparable to this study, previous studies had indicated lane-use values within this range. Longer auxiliary lanes and lower downstream right-turn volumes encouraged the use of auxiliary lanes by straight-through vehicles. The range of lane utilization adjustment factors $(f_{\mathrm{LU}}$-factors) calculated from field data was 0.72–0.80, which is lower than the 1997 U.S. Highway Capacity Manual default value of 0.91 for a three-lane through/right-turn group.

Asphalt pavements under low temperature conditions develop transverse cracks as a result of one or a combination of three distress mechanisms: (1) single-event thermal cracking; (2) thermal fatigue; and (3) load-associated thermal cracking. The cracking temperature of asphalt pavements based on the single-event thermal cracking mechanism of the asphalt binder has been suggested as a good measure of the low temperature criterion for use as a binder purchase specification. The determination of the single-event cracking temperature $T_{\mathrm{cr}}$ requires a comparison of the pavement thermal stress computed from the binder stiffness data generated from the bending beam rheometer (BBR) with the binder strength data measured using the direct tension test (DTT). Thus, two instruments, each with different sample preparation techniques, different sample sizes, and different data analysis methods, are required for the final determination of a single point signifying the cracking temperature. The present work revisits the method of data analysis to determine $T_{\mathrm{cr}}$ and shows that the cracking temperature can be obtained directly from the BBR data without the use of the DTT. The procedure uses the pavement thermal stress computed from the BBR binder stiffness data to calculate the intersection of the two asymptotic thermal stress buildup rates, which directly identifies the cracking temperature. This procedure has a significant advantage in that it makes the BBR a stand-alone device for the determination of the low temperature binder specifications.

Many studies conducted over the last 20 years show that speed bumps have only a limited affect in decreasing traffic speed, especially compared to better designed passive speed control devices, such as speed humps or cushions. These findings were confirmed during an investigation described in the present study conducted in an Italian town. Besides the well known phenomena, such as sudden slow down and acceleration before and after the bump (which is often a source of accidents), it was observed that drivers tend to perform all sorts of maneuvers to reduce their discomfort as much as possible when approaching the undulation. This fact is particularly evident on streets where the speed bumps do not extend over the entire street. This kind of misplacement seems to encourage drivers to perform avoidance maneuvers, such as short deviations into the opposite lanes or parking lanes and bus stops. This study is an attempt to reach a qualitative and quantitative classification of possible alterations in drivers’ behavior due to the presence of the speed bumps. The result of more than 25,000 observations of cars and motorcycles reveal that speed bumps, where not properly installed, induce many drivers (up to 50% for cars and up to 85% for motorcyclists) to avoid the undulation to reduce the level of noise and vibrations perceived inside/on the vehicle.