In the present paper, a computer vision system composed of five models is proposed for vehicle classification. The five models are: (1) The perspective projection model; (2) the length measurement model; (3) the width and height estimation model; (4) the profile character extraction model; and (5) the tree‐type classification model. As a significant feature, the third model provides an effective way to estimate width and height of vehicles from video images. This capability is not available in loop detector systems at present. The fourth model provides an approach to obtain two important profile characters of vehicle namely: (1) The front shape of a vehicle: flat front of projecting front; and (2) the number of units of which a vehicle is composed. With these characters, not only can buses be differentiated from trucks, but also vans can be separated from cars. Therefore, common drawbacks of present vehicle‐classification systems are remedied. Experimental results and suggestions for the future improvements are also discussed.

The automation of urban freeways is a major feature of the new surface transportation act. Reducing freeway congestion can be presumed to have drawbacks, however, in terms of induced travel and possibly greater travel costs and higher emissions. We test several freeway‐automation scenarios for effects on travel demand using a regional four‐step model set. A partial‐automation scenario has the lowest congestion, but higher vehicle miles than conventional scenarios (transit expansion, new high‐occupancy vehicle (HOV) lanes, doing nothing). We projected emissions with the official California models and found that the conventional scenarios produced lower emissions for most pollutants. Using our own traveler‐cost model, we found the same general rankings, with all the automation alternatives more costly than transit, HOV lanes, and doing nothing. These results should be seen as tentative and be used to guide further research. Improved automation scenarios and more accurate methods of projecting travel demand, emissions, and costs are outlined.

The design of transportation‐network service routes is both combinatorially complex and fraught with multiple and conflicting management objectives. The focus of this work is the design and development of a computer‐aided system for assisting in the design of network service routes. The system integrates: (1) Spatial network data; (2) multiobjective heuristic optimization techniques; and (3) an interactive, user‐controlled graphical interface. The result is a systematic methodology for evaluating route configurations that facilitates the identification of solutions that reduce service‐resource requirements while at the same time improve the overall level of service to the network. Although this research was motivated by the need to design winter snow‐ and ice‐control service routes in Indiana, the approach is generalizable to a wide variety of applications requiring service to the edges (road segments) of a network such as crack sealing, pothole repair, painting and striping, weed control, and scheduled inspection. The structure and function of this system are discussed, as well as an evaluation of its use to date by Indiana Department of Transportation maintenance engineers.

The low‐floor bus design is generally recognized as an effective tool for serving the population of transit passengers with disabilities. This design, in which the main platform is reached with a single step from curb level, is evaluated here for its potential to reduce dwell times of buses in regular service with regard to the nondisabled population as well. Three data sources were developed in cooperation with the Ann Arbor (Michigan) Transportation Authority; a regular service sample in which the full range of fare‐payment media was used, a special shuttle sample, in which all fares were prepaid, and a limited wheelchair trial. Descriptive analysis of the data as well as multiple regression modeling indicate dwell‐time savings accruing to the low‐floor bus design overall. Ranges of potential savings by passenger type and fare payment media are also estimated. The study concludes that exclusive of benefits in wheelchair passenger boarding times, the low‐floor design may shorten general passenger boarding and alighting time by 13–15%.

The transportation‐planning process incorporates the development of representative‐trip‐generation models from base year data for making forecasts of travel in subsequent years. Category analysis is a technique based on the unit of the household and its characteristics. The data are cross‐classified over various socioeconomic schemes. Trip rates based on such schemes are made for a particular category of analysis unit. When the household characteristics are cross‐classified with the available data, certain cross‐classification cells may get sparse data and certain others may not get any data. Under such circumstances, conventional application of category analysis may not be appropriate. To overcome this difficulty a linear statistical model framework was developed. The estimated mean trip rates for the cells of the cross‐classified table use a model fit based on data from all cells and not just the data from any given cell, which may not be reliable. Developed models were illustrated for the different cross‐classification schemes.

A review of several relationships commonly used by pavement engineers for structural characterization of subgrade soils, subbase, and base is presented. A careful examination of the literature on the subject revealed that some of the cross‐correlations presented in the 1986 American Association of State Highway and Transportations Officials (AASHTO) thickness design guide (Guide for Design of Pavement Structures) are either invalid or inaccurate. Problems associated with many of these correlations are put in the context of their historical development. A clarification is provided to the current inaccurate cross‐correlation information reported in the 1986 AASHTO guide with regard to the Kentucky version of California bearing ratio (CBR). A common scale for correlating “soil‐strength/ stiffness” parameters, including the resilent modulus, is presented for use with the AASHTO guide as well as mechanistic pavement‐design procedures. A refinement of the AASHTO procedure for selection of resilient modulus of granular base and subbase materials is also presented. An analysis of issues related to measurement of resilient modulus indicates that there is a great need for standardization of equipment and test procedures. In this regard, some of the issues related to the field versus laboratory measurement of resilient modulus are discussed.

Quality of highway construction is a major concern for highway agency engineers and contractors, and density of asphalt‐concrete pavement is one of the primary quality‐control parameters. The present study evaluates the use of the nuclear gauge as a prospective quality‐control tool by comparing nuclear‐gauge performance with core density measurements. Density measurements were recorded by the contractor and the highway department using both core and nuclear‐gauge methods on four paving projects, and statistical analyses conducted to compare the results. The analyses indicate that there is no statistical difference between the testers (contractor or highway department) or in the variability of the measurements obtained by the different methods. However, it is observed that the means of the measurements obtained by the different methods are statistically different and that density variability depends on type of mix. A multiple‐regression model is developed that correlates core density with nuclear gauge readings and type of pavement.

The six most commonly encountered aggregates that are used as subbases/bases of roadways in Oklahoma were selected and tested under cyclic loading to evaluate the Resilient Modulus (RM) by using the AASHTO T292‐91I and T294‐92I methods. The gradation of the specimens met the Oklahoma Department of Transportation 1988 specifications for type A (fine) and type B (coarse) materials. A series of exploratory tests were performed to assess the effect of gradation, compaction method, specimen size, and testing procedures on the RM values. The values reported or suggested by various agencies were compared with the values obtained from the present study. Statistical correlations were established between RM and the California bearing ratio, and between RM and cohesion and friction angle. It was observed that, for a given gradation and bulk stress, the RM values obtained in this study were slightly lower than those reported in the literature; the influence of gradation and compaction method on RM values were less significant compared to the effects of the stress state. Higher RM values were obtained from the T294‐92I testing procedure than those obtained by the T292‐91I testing procedure. The RM values for 10.16‐cm specimens were higher than those for 15.24‐cm specimens, and a satisfactory correlation did exist between the RM values and the cohesion and friction angle.

The present paper is concerned with estimating the reliability of underground pipelines subject to externally applied loading and to corrosion, either externally or internally, or both. This affects the strength of the pipeline in resisting the applied loads so that the reliability would be expected to decrease with time. Corrosion is a phenomenon about which there is a great deal of uncertainty. One way of allowing for this is through probabilistic modeling of the material loss as a nonlinear function of time. This approach is explored in the present paper, using a nonlinear function first postulated for atmospheric corrosion. This model is incorporated into expressions for stress resulting from externally applied loading and internal pressure to provide a limit state function expressing the boundary between survival and failure of the system. The probability of system failure is then evaluated using techniques developed for structural reliability analysis.