Applications of Advanced Technologies in Transportation (2002)

The University of Massachusetts Amherst (UMass) and the Massachusetts Highway Department (MassHighway) have entered into a partnership to develop, implement, and operate a regional traveler information center (RTIC) for the Pioneer Valley of Western Massachusetts. The project is unique because UMass will offer operational support for MassHighway's regional traveler information system, in addition to collecting, processing, and disseminating real-time traffic information to the traveling public and regional transportation providers. RTIC will maximize the quality of traffic data available by integrating data and information from multiple sources, and providing value-added information such as estimated travel times and suggested incident response plans. Because RTIC is based at a major research university, it will also provide a real-world test-bed for the development and evaluation of advanced traffic technology, software, and traffic management procedures. RTIC is a cornerstone for coordinated ITS deployment in the region, and leverages technical and financial resources that might otherwise be unavailable in a rural setting. It is envisioned that the system will gradually expand to cover other rural areas of the state, and integrate with other wide-area traveler information systems under development in the Northeast.

Traffic assignment has received much attention since 1950s'. Stochasticity, dynamic and fuzziness are three characteristics of traffic assignment. Since drivers always have no perfect information about traffic conditions, they choose their route based on perceived travel time. Traditional traffic assignment model can only capture the stochasticity and dynamic features of traffic assignment and cannot represent the fuzziness of driver's perception over travel time. Thus, we present a fuzzy dynamic traffic assignment (FDTA) model in this paper. By the definition of fuzzy perceived link travel time and fuzzy perceived path travel time, we use fuzzy shortest path algorithm to find the group of fuzzy shortest paths and assign traffic to each of them proportionally to their memberships. The proposed FDTA model can be used in the real-time network load management, especially with variable message signs.

The variability of travel times on a given route for a particular journey is known to be an important factor in influencing the choice of mode and route for trips, particularly for those trip types where constraints may exist on arrival time at the destination. Previous work in this area has concentrated on long-term variability, such as daily and seasonal variability. This paper describes some results obtained from a study of short-term travel time variability. Data generated by simulation permits the study of the distributions of short-term travel times and the correlation among link/route travel times. It was found, for instance, that strong evidence of significant correlation between link travel times and between link travel time and arrival time, and that conditional link travel time distributions could be represented by normal distributions. Some possible extensions are suggested.

Predictive route guidance uses forecasts of traffic conditions to provide drivers with messages related to their route choice decisions. Three distinct approaches for generating predictive guidance — extrapolation methods, methods based on traffic network models, and predictive feedback methods — are presented and discussed.

This paper introduces a model to estimate average link travel time of signalized arterials using data obtained from detectors in Sydney Co-Ordinated Adaptive Traffic System (SCATS). The proposed model makes use of detectors at the upstream stop line of a link to capture the arrival of vehicles, and use a platoon dispersion model to project the vehicle arrival times at the downstream stop line. Using the vehicle departure patterns obtained at detectors at the downstream stop line, the average link travel time is then estimated. The accuracy of the proposed model has been tested, by means of microscopic traffic simulations, for link lengths between 300 m to 800 m, and at different urban street level of service. Results indicate that the proposed link travel time estimation model has an average absolute error of 2.26 sec.

From an operations standpoint, the most important task of a traffic surveillance system is determining reliably whether the facility is free flowing or congested. The second most important task is responding rapidly when the facility becomes congested. These tasks are complicated by the fact that conventional vehicle detectors only monitor discrete points along the roadway. The detectors are typically placed at least 0.5 km apart and conditions between the detectors must be inferred from the local measurements. It can take several minutes before an incident between detectors is observable in the point measurements and even longer to differentiate between noise and a true incident. To address such issues, this paper uses existing detector stations to match vehicle measurements between stations and monitor the entire roadway. Earlier work has already demonstrated the methodology using effective vehicle lengths measured with dual loop detectors. This paper extends the work by presenting a method to estimate vehicle lengths from single loop detectors and then applies the vehicle reidentification methodology to these common detectors with good results. The work should prove beneficial for traffic management and traveler information applications, while promising to be deployable in the short term.

This paper describes research investigating corridor travel time mean and variance estimation. Simultaneous travel time data were collected along two corridors of approximately 2.0 miles (3.2 kilometers) each. Travel time data from test vehicles instrumented with distance measuring instruments (DMIs), commercial vehicles by video, and intelligent transportation systems (ITS) data sources were collected. Link and corridor travel times were estimated from the ITS data source (either AVI or inductance loop detectors) and compared with travel times measured from instrumented test vehicles and commercial vehicles. Several important findings are described in this paper. AVI was found to be a cost-effective travel time data collection method based on a comparison of travel time estimates with estimates obtained from instrumented test vehicles. A procedure was developed to identify which drivers are in need of additional training on the test vehicle data collection technique. Commercial vehicles were found to have longer travel times than those reported by the ITS data source along the two study corridors. Finally, an empirical analysis is provided that identifies corridor travel time mean and variance estimates from link travel time mean and variance estimates that make up the entire corridor. The correlation between adjacent links along a corridor was investigated with empirical ITS travel time data from each site. Finally, the difficulty of using spot-speed measurements (i.e., estimates from loop detectors) for corridor travel time variance estimates is also identified.

The continuous increasing traffic volume within the recent years leads to novel appearances of traffic characteristics. Clearly, a proper way of modeling these conditions is prerequisite for an adequate design of simulation and prediction tools. Although the permanent rising use of vehicles often implies a very high degree of bursty traffic behavior, great efforts are still taken to adapt the implemented ARMA and Markovian prediction tools to the current situation. However, theses models possess the disadvantage of smoothing out turbulent traffic due to its finite memory. In contrary, for high traffic loads long-range dependent (LRD) processes have proved to be the superior, adequate alternative in similar other fields like electronic network traffic (e.g. TCP/IP and ethernet traffic), video-on-demand streaming, and stock market prediction. Thus, this paper first gives a precise, unique understanding into the mathematical background of LRD processes. Based upon this knowledge, appropriate, economic implementation techniques for real traffic control systems will be developed. Particular attention is laid on two major points: first, economic constraints require cost-efficient solutions for already existing systems. So the mainly used, ARMA/Markovian tools will be upgraded to LRD systems by the expansion to long-memory FARIMA processes. Second, importance has to be laid to obtain a clear relation between abstract, mathematical parameters and existing boundary conditions. Here, the superposition of heavy-tailed distributions, each of them representing the local traffic condition of individual roads, gateway entries etc. leads to the desired Fractional Brownian Motion. Therefore, turbulent conditions, e.g. on highway corridors, can be depicted by the result of its incoming traffic.

The paper introduces a methodology that uses global positioning system (GPS) data to calibrate simulated vehicle speed along signalized arterials in PARAMICS, a microscopic traffic simulation model. The objective of this calibration is to minimize the absolute error between speed profiles along an arterial obtained from GPS and simulated data, under different signal phases at both ends of the link. The calibrated parameters in PARAMICS were link speed limit, maximum acceleration rate, mean target headway, mean reaction time and minimum gap. The calibration has enabled PARAMICS to generate speed profiles with an average absolute error of not more than 5.0 km/h, which is at least an 11% improvement from using the default parameter values.

In this paper, the origin and basic theory of traffic impact fee are reviewed, and the unjustness of American policy of traffic impact fee and the principles of cost allocation are discussed. Furthermore, allocation model of traffic impact fee and corresponding algorithms are put forward, and the effectiveness of them is summarized.

The Massachusetts Bay Transportation Authority (MBTA) is the transit authority for the Boston region providing commuter rail, rail rapid transit, commuter boat, paratransit, and bus service. A current policy goal at the MBTA is to develop an improved fixed-route bus system. To accomplish this goal investment in the bus fleet is underway and inclusion of ITS features in the investment mix is key. Two recent contracts designed to improve bus service include one to retrofit 95 diesel buses with GPS based automatic passenger announcement systems and another to introduce a number of ITS features on a new bus rapid transit (BRT) system. The BRT system will provide wayside, central control, and Internet information on bus arrivals, as well as, the trigger for traffic signal priority and is scheduled to initiate service in spring 2002. The MBTA also anticipates the delivery of more than 400 new vehicles, some to replace retired buses, and others to operate on new services. Many of these vehicles will be CNG powered. Delivery is expected in 2003–2004. In addition to the new fuel type, the MBTA will also be outfitting all of these buses with GPS based automatic passenger announcement systems, and linking some of them to the central control with a computer aided dispatch system. The implementation of the new ITS technology will help make the goal of improved bus service a reality. However, successfully integrating the new infrastructure into an existing organization requires knowledge of both where the MBTA is and where the MBTA wants to go.

Electronic Toll Collection (ETC) has become a technical feasible and financial viable alternative for financing and maintaining highway construction and operation. In United States, many toll road operating agencies have successfully used Intelligent Transportation System (ITS) technology to administer planning, design, operation, management, and financing of highway facilities, and to improve the inherently labor-intensive toll collection process. This study examined the effective application of ETC technology for the planning and design applications on freeways and expressways in Taiwan, Republic of China (ROC). This paper reviewed various design and operational issues, including the current application of AVI/TOLL technologies, potential operational enhancements, and alternative design upgrades, for effective Electronic Toll Collection (ETC) and Electronic Toll and Traffic Management (ETTM).

In order to accurately evaluate the impact of the implementation of Intelligent Transportation Systems (ITS) services such as route guidance and Advanced Traveler Information Systems (ATIS), a Simultaneous Route and Departure time choice (SRD) model is needed. This paper develops a cell-based SRD formulation for the case of elastic total demand through the Variational Inequality (VI) approach. This formulation follows the ideal Dynamic User Optimal (DUO) route/departure time choice principle and the First-In-First-Out (FIFO) conditions. It encapsulates a network version of the Cell Transmission Model (CTM) to improve the accuracy of dynamic traffic modeling. To solve the formulation, we adopt a descent method developed for co-coercive variational inequality problems. In the numerical study, we solve a 13-node, 19-link network to demonstrate that the solution of this formulation satisfies the ideal DUO route/departure time choice principle precisely.

This study demonstrates a strategy for identifying bottlenecks on a network of roadways containing toll facilities. The network is divided into 295 highway segments, 58 of which contain toll facilities. It successfully implements TNCC, the Toll Network Capacity Calculator methodology, to calculate capacities of the 58 highway segments containing toll facilities. The capacities of the remaining 237 highway segments were obtained by the use of traditional capacity calculation methods borrowed from the Highway Capacity Manual 2000. Observed volumes and comparison checks of the Plazas' capacities with the capacities of the highway segments prior to the toll facilities supports the identification of bottlenecks on the network.

Customer input in the planning, design, and deployment of Advanced Traveler Information Systems (ATIS) is highly desirable in ensuring its success in serving the needs of travelers. A systematic method by which customer preferences and opinions may be incorporated into the process is proposed in this paper. The authors develop a Customer-Integrated ATIS Planning and Design process which includes goal setting, target market(s) selection, identification of customer needs and preferences, and selection of ATIS media and content; these steps culminate in a prototype of the ATIS. The Customer-Integrated ATIS Deployment process proposed by the authors consist of the trial deployment of the ATIS prototype, customer trials and feedback, modifications to the media and content, and the full deployment of the ATIS, with continuous customer feedback. The specifics of both processes are discussed in detail in this paper.

At the Luis Munoz Marin International Airport in San Juan, Puerto Rico, the Puerto Rico Ports Authority is developing a dual midfield / cross-field taxi-way. The new taxiway is part of an ongoing expansion program at the airport that will provide bi-directional traffic between the airport's two runways. The dual taxiway system separates the terminal facility from the other major operational areas on the airfield and intersects the single "at-grade" service road that connects the two areas. This road provides access between the terminal and operational areas that house vital services such as cargo, baggage handling, catering, and Airport Rescue and Fire Fighting (ARFF) facility. The paper describes the application of intelligent traffic control system components to provide redundant control of service vehicle crossing traffic at the taxiway/service road intersection while providing safe separation from aircraft operating on the dual taxiway system. The new control system serves a dual function of enhancing safety by preventing incursions between aircraft and ground service vehicles as well as optimizing traffic flow for service vehicle crossings.

In conducting a "best practices and technologies" study for the deployment of ITS technologies in Ohio, the Ohio DOT determined that the study should focus on the macro corridor system in the state. This paper focuses on a summary report of a major research report submitted to the Ohio DOT in mid-2001 that explored the causes of delay on these macro corridor highways and suggested those management practices and ITS technologies that would be most effective for Ohio's ITS program. In terms of congestion on highways in the three largest metropolitan areas in Ohio, the percentage of congestion caused by incidents was approximately 55 percent in the most recent Texas Transportation Institute rating. The ITS program recommended for further deployment to the Ohio DOT included (1) non-technical freeway operations improvements, (2) synergistic collection and analysis of traffic data, (3) multi-agency traffic management with DOTs at the state and local level, police and fire, (4) incident detection, (5) video surveillance, (6) traffic control during incidents, and (7) traveler information dissemination. A detailed assessment of various technologies in each of these seven functional areas was made with an array of state-of-the-practice technologies, but no one specific technology was recommended for statewide application. In this way the ITS partnerships that are formed by the various ODOT districts would have information to guide them in their choices of technologies. In terms of management practices, it was noted that Ohio's highway and multimodal transportation program is subject to the oversight and policy guidance of a Transportation Review and Advisory Committee (TRAC). It was recommended that ITS projects be eliminated from the TRAC review process, or that ODOT develop criteria for ranking and selecting ITS projects based on their value in improving operations on Ohio's macro highway system.

Traffic congestion is a major problem that is becoming exceedingly difficult to address using conventional transportation and traffic engineering tools. Consideration of the characteristics of the Beirut, Lebanon network have led us to identify the management of incidents as a primary area of immediate intervention, where ITS-assisted technologies could provide significant payoff. This research develops a framework for assessing the deployment of incident management technologies, and applies the framework in the context of the Greater Beirut Area (GBA).

Data collection and data processing determines the success of an incident detection system. The former provides the traffic data types and affects the configuration and performance of the resulting data processing algorithms. There have been a countless number of methods and algorithms developed for incident detection. They mainly rely on inductive loop detectors and loop emulators. Point data sources are notoriously unreliable and do not provide the best traffic data for determining real-time traffic conditions. The results of previous research indicate that no significant improvement on incident detection performance will be obtained until new types of traffic data collection technologies with high accuracy, reliability and coverage are used. This paper identifies and defines a new classification system based on three distinct incident detection types: roadway-based, probe-based and driver-based, in terms of their different sensor data (collection) types. The results of our nationwide survey on traffic management centers and traffic operations centers about their current and soon to be implemented traffic sensors and incident detection strategies suggest a movement from the traditionally used traffic sensing techniques to driver-based witness report. No current traffic sensing technology can provide the richness of incident detection and description and the coverage on roadway systems like driver-based incident detection techniques. In this paper, driver-based techniques, their effectiveness and performances are reviewed, identified and assessed. The implementation and benefit-cost issues are also analyzed and briefly compared with other incident detection approaches.

This paper aims to develop a model for an incident response time on freeways. This study utilizes a ten-year crash-data of the state of Ohio from 1990 to 1999. Incident response times on freeways were found by using crash log entries of crash-report time and crash-scene-arrival time in the crash data files. The generalized form of Poisson regression model (Negative Binomial regression model) was eventually developed. However, the additional data that affect the capacity of freeways such as number of lanes and traffic demand at which an incident occurred will help develop accurate model.

This paper introduces a feedback simulation-based approach to evaluating freeway diversion route plans in integrated incident management systems using real-time data. The proposed approach employs an anticipatory technique to estimate demand and incident severity based on real-time data transmitted to the TMC through a fiber optic communication network and a library of historical traffic volume and incident data. Using the anticipated volume and incident data in a microscopic simulation model for the freeway and arterial systems network, an optimal decision about freeway diversion plans can initially be reached. Using real-time freeway and exit-ramp volume data, the initial plans can be reevaluated and readjusted accordingly. The proposed approach is intended to be used as a decision-making tool for real time incident management, specifically for developing and testing diversion plans using real-time data. Ten incidents with different severity levels and durations were used as a case study for the evaluation. Results from the analysis showed that the proposed evaluation framework produced optimal diversion route plans. The benefit of using real-time operational characteristics in the evaluation seems to be greater for incidents of moderate severity and duration. The results showed that the proposed approach can be used as a decision-making tool for real-time incident management, specifically for developing and testing diversion plans using real-time data.

This paper presents incident prediction models for the I-80/94 (Borman Expressway in northwestern Indiana) and I-465 (northeastern Indianapolis) freeway sections. The primary independent variables are traffic volume, truck percentage and weather. The models were fitted using two alternative methods, non-linear regression and maximum likelihood methods. The results, which were similar for both methods, showed that volume unlike length is non-linearly related to incidents and therefore volume and section length have to be considered as separate variables in the modeling process. Truck percentage was one of the significant factors that determines the number of incidents. The results from the non-linear regression and maximum likelihood estimation methods were found to be similar.

With the rapid increase of motor vehicles, the increasing traffic jams hinder the development of most cities in China. Based on the abundant practical experience on traffic management and the large amount of practical data in Gourd Island city, this paper synthetically analyzes the reasons for traffic jams, which are not only related to the contradiction between traffic needs and traffic offerings, but also related to the unreasonable distribution of traffic generations, irrelevant structure of transport capacity and road network, insufficiency in public traffic capacity and means of transportation as well as lagging traffic management. Furthermore, the possible ways to alleviate traffic jams in Gourd Island city are discussed in details. Finally, the project of urban traffic integrated planning in Gourd Island city is proposed so as to solve the traffic problems radically.

The paper presents space occupancy (S_oc) as a measure of state of traffic congestion in real time. To accurately measure state of traffic, space occupancy (S_oc) is calculated as the percentage length of highway occupied by vehicles, taking into account length of every vehicle. The paper presents the concepts of S_oc and describes the usefulness of utilizing S_oc compared to density. The relationship between S_oc and density is also determined and a traffic simulation model CELLSIM is used to evaluate the accuracy of each parameter. The parameters were studied using different length of vehicles and mix of traffic. S_oc more accurately determines the magnitude of congestion and may be used to set accurate thresholds of traffic congestion. The concepts described can be applied to incident detection in real time control of network traffic flow in Intelligent Transportation Systems.

This paper presents the application of Parallel Genetic Algorithms (GA) on signal coordination for networks with congested intersections. When a serial GA is applied to solve traffic control problems, its performance in terms of computation time diminishes as the size of signal networks increases, or the duration of congestion lengthens. This paper uses master-slave parallel micro GA (PMGA) to solve signal coordination, formulated as a dynamic optimization problem. The elapsed time per generation reduces as the number of processors involved increases. When the number of processors equals the number of individuals, for the network described in this paper, the master-slave PMGA reaches the highest speed-up. Although the master-slave PMGA provides a lower bound speed-up and its efficiency degrades as the number of processors increases, this study demonstrates the potential benefits of using parallel GA in solving signal coordination problems.

This paper presents a lane-based optimization method for the integrated design of lane markings and signal settings for isolated junctions, in which the integer variables include the permitted movements on traffic lanes and successor functions to govern the order of signal displays, whereas the continuous variables include the assigned lane flows, common flow multiplier, cycle length, and starts and durations of green for traffic lanes. A set of constraints is set up to ensure feasibility and safety of the resulting optimal lane markings and signal settings. The delay minimization problem for fixed-time traffic signal at isolated junctions is solved by two heuristic algorithms. A numerical example is used to test the effectiveness of these algorithms.

The law enforcement arm for a growing number of agencies is turning to automated camera technology for their enforcement of the red light violation. Traffic signal agencies are often called upon to assist in determining the feasibility and placement of automated enforcement. While there are systems on the market that can perform data collection of RLR occurrences, the City of Franklin, Tennessee has found it more feasible to gather this data using its own resources. This paper presents a method for traffic signal agencies to collect this data as an aid to their analysis of RLR activity. The collection method consists of overhead videotaping of selected intersection approaches during key periods of traffic activity. The videotape recording is then returned to the office for playback and tabulation of the approach activity for the period. The tabulation form is designed to enable the identification, for each turning movement, of the frequency of vehicles passing through the green, yellow or red display. The paper describes four ways in which the RLR videotape sample serves the agency. The paper also discusses the design of the tabulation form, the various data items that it makes available, the issues encountered in its application, and the experience gained in the use of videotaping for data collection. The paper finally describes the analysis of the data, as an aid to the City's RLR camera feasibility investigation. The author submits that other agencies will find the method useful, and offers the form as a possible standard for future publications on traffic engineering studies.

Freight movement in the United States is severely crippled by congested roadways leading into and out of port areas. This congestion is due, in part, to significant growth in freight movement over the last decade. In the next twenty year, international marine trade is expected to double, pointing to the need for improvements in landside access to port terminals. Technological advances made in the movement of freight at port terminals may provide opportunities for using valuable information available through these advanced technologies for use in optimizing and controlling freight movement on roadways accessing the port area. Information available through advanced technologies and freight management systems can be integrated with roadway management systems to promote an integraded dynamic information system for freight movement. This paper provides an approach for accomplishing this integration using control theory.

One of the key elements for an effective ATMIS is the ability to predict traffic conditions with reasonable accuracy such that transportation system users are able to make better travel decisions. This study evaluates and implements a traffic prediction model on I-4. The performance is evaluated against different prediction horizons and traffic conditions. The algorithm was also used to support the implementation of an online traffic prediction system that provides the freeway travelers with predicted travel times and delays along the freeway corridor.

Driving simulation provides a vastly improved mechanism for conducting driver behavior and comprehension research. Presently, at least 40 known driving simulators are located at research institutes throughout the world. Although research conducted in the actual driving environment is considered to be optimal, the use of a driving simulator allows for multiple variables and scenarios to be evaluated in a timely and cost effective manner without losing the field credibility. Driving simulation places drivers in a fully interactive dynamic scenario just as if they were actually driving. This paper describes a driver behavior analysis conducted using a full-scale dynamic driving simulator as a tool for evaluating driver comprehension of protected/permitted left-turn (PPLT) signal displays.

One of the key strategic principles of the National Military Strategy of the United States is crisis response through power projection, which requires rapid mobilization to worldwide situations. Power projection is the ability to apply all or some of the instruments of national power, political, economical, informational and military. Force Projection is the military instrument of power projection and its primary purpose is to deter threats to U.S. interests. Force Projection is the deployment, sustainment, employment and redeployment of military force from Continental United States. The need for Force Projection came about after the end of the Cold War and with the rise in political, economic and military capabilities of regional powers. With increase in hostilities to the United States and its vital global interests, deterrence can only be accomplished through the ability to rapidly project forces.

This paper will discuss the U.S. Army program to replace Halon 1301 fire suppression systems used onboard Army watercraft and the U.S. Department of Transportation, Volpe National Transportation Systems Center's (Volpe Center) approach to the identification, research and testing, selection, design development, and installation of a prototype alternative system. The paper includes a review of the existing/emerging state-of-the-art technologies, an account of user acceptance of the new system, lessons learned, and possible future research needs.

This paper presents initial results of the air quality analysis project that is focused on remote sensing laser measurements of air pollutants and air quality impact of transportation systems. Tunable laser remote sensing technology is being used over a distance of several hundred meters for measuring Nitrogen Dioxide and Ozone related to transportation. The laser is tuned between ultraviolet, visible, near infrared, and thermal infrared spectral regions. The difference in the absorption of light at these different wavelengths is used to determine concentration of air pollutants. The absorption band of Nitrogen Dioxide lies in the visible part of the electromagnetic spectrum. It absorbs blue color of sky, resulting in brown color smog. The laser measurement results of Nitrogen Dioxide concentration, measured near Mississippi Highway 6 in Oxford, showed nearly 25 times more concentration at 10 AM (daytime) compared to the measurement at 11 PM (nighttime) when the traffic was minimal. Higher Nitrogen Dioxide concentration is also associated with higher air temperature in the daytime measurement.

Intermodal transportation activities have seen fundamental changes in technology, operation, management and organization in the past three decades. A transportation carrier deliver products to its clients using containers in various transportation modes such as railcars and trucks. After delivery and unloading by the customers, the empty containers are shipped back to a depot or warehouse designated by the carrier for subsequent shipment to other customers. After loading the new customer's products, the containers are transported directly to their destinations or through some intermediate depots. Under this type of logistic networks, containers spend a significant amount of time in empty movements. For example in the US rail system it is estimated that a railcar is empty during 40% of its average car cycle. The importance of empty container movement in carriers cost structure has inspired a number of studies directed at managing empty container fleets for railroad, liner shipping operators, and truckers. A fundamental questions these studies address is: Given an inland logistic network, how should the carrier dispatch empty containers to meet shippers' demand, to relocate empty containers among depots and warehouses, and to lease on/off containers in preparation for future demand. While these studies have achieved some degree of success in reducing empty container related operational costs, a more fundamental and strategic question is how to design the underlying logistics network; namely, how to determine the location and capacity of depots and warehouses in order to achieve a broader and more significant cost reduction. This paper addresses the above question by developing a decision support tool to assist transportation carriers in determining their inland logistic network, the location and the capacity of each depot.

The existing formulae for determining computed width of out-of-gauge goods transported by railway can not reflect the real out-of-gauge degree. In this paper, theoretical basis for determining computed width is discussed and new formulae which can form reliable basis for determining transport condition of cars loaded with out-of-gauge goods are put forward. The paper will play an important theoretical and actual role to improve out-of-gauge goods transport theory.

As we enter the twenty first century, humankind is blessed with unprecedented mobility. Global circumnavigation has become a matter of hours, compared with the proverbial 80 days of Jules Verne. Improvements in rural transport have lowered the costs of agricultural production; directly, by increasing access to markets and credit, and indirectly by helping to develop the nonagricultural rural economy. Improvements in urban transport have increased labor market mobility and access to employment opportunities while changing the scale and form of urban agglomerations. Improvements in inter city and international transport have facilitated commerce, trade, tourism and supported the globalization of industry and services. But the blessings of good transport networks and services have not beneficiated mankind uniformly. Peak hour traffic congestion is worsening, especially in the suburbs, exacting an economic toll in lost time and productivity. Travel in single occupancy vehicles is the greatest contributor to urban air pollution, which poses health risks for millions. It is also the single largest factor in the buildup of greenhouse gases in the atmosphere, which risks causing climate changes and global warming. One of the most contentious issues is the impact of mobility—on forests and other ecological habitats and cultural sites. Long-distance trucking is a major source of transmission of communicable diseases, particularly HIV/AIDS, in developing countries. With the augmented global mobility, these fallacies seem to engraving the humanity. Being blessed with the inventions of information technology, the traditional pattern of mobility is sure to change. E-commerce, E-money has already become the dominant business mode in the western countries. E-travel is sure to flourish. In order to acclimatize in the coming future; the blessings from virtual traveling must be winnowed to ameliorate the condition of the people of the world.

In this paper, both the consciousness of, and the actual conditions of, a cooperative pick-up and delivery system for goods in the Tenjin District of Fukuoka City, Japan, are investigated from the viewpoints of the shippers and the transportation companies. Details of survey include characteristics of shippers, office, goods, pick-up and delivery. The factors for the use and participation in the cooperative pick-up and delivery system are analyzed using questionnaire data. Effective strategies to use such a system to participate in it are also discussed by considering the actual conditions. As a result, PR on the social merits of the system to employers for shippers and to encourage participation of the small-scale shipping agencies for transportation companies are necessary.

This paper outlines the software for a modernized version of dial-a-ride called ADART. It is a sequel to the original concept paper, which addresses service costs and benefits from the point of view of the operator and customer. In contrast, this paper maps ADART's features and operation into software, describing the functionality of it major subsystems in order to provide a brief overview of how the system works from the perspective of an engineer.

Different kinds of trains have been operating on busy railway lines of China. Traditionally in monopoly state, railway operators were not willing to offer passenger train service with higher commercial speed due to line capacity. However, railway operators have been forced to increase the passenger train speed in order to keep their market share in competition with highway transport which has been developed greatly since 1980's in China. This paper analyzes the factors related to the feasibility to increase passenger train speed on Chinese railway mainlines. A target-speed based control policy with technical and operational constraints has been advanced for the purpose to achieve higher speed for passenger trains. With train movement simulator developed by authors, some simulation results have been got to study the feasibility of increasing passenger train speed on mainlines with little reform on existing infrastructures. The research may offer a reference for Chinese railway operators to work out new train schedule.

The planning Beijing-Shanghai high-speed railway (Jing-Hu HSR) in China will adopt mixed operation of high-speed trains (HST) and medium-speed trains (MST), which greatly complicates the transportation management on this line. A synthetic simulation system is developed to study the transportation management under this special operation mode. The simulation system aims at not only making train operation plans for various operation conditions, but also providing a good experimental tool to evaluate the synthetic performance of transportation system from both static and dynamic aspects. To achieve the goal, the basic structure and data flow of the system are designed considering the whole process of train operation from making plan to executing plan. Some key technologies to guarantee the optimal simulation results are simply described. By use of the system, a lot of simulations on train operation are made and some valuable conclusions and suggestions are proposed.

China has very wide territory with more than one billion in population. Its fast developing economy will reach to the level of economies of middle-developed countries in the middle of 21st century. The needs for high-speed passenger transportation are increasing quite rapidly. This paper discusses necessary development strategies for high-speed passenger transportation in the coming years. The paper also summarizes the progress of a-high-speed MAGLEV project in Shanghai, China.

The National Airspace System Resource (NASR) System provides the Federal Aviation Administration (FAA) with the means for storing and maintaining a reference database with descriptive details of the National Airspace System's (NAS's) infrastructure and the operational status of all components. NASR, completed in January 1999 and declared operational in March 1999, assures the FAA of current and accurate information available on the framework upon which Air Traffic Control and other FAA activities are overlaid. NASR supports the management of NAS data used by the FAA, chart producers, military and other Government entities to produce a variety of aeronautical publications. In addition, NASR data is accessed by commercial air carriers, flight planners, and flight systems developers as required by FAA regulation to support the safety-of-flight mission. NASR is a client-server system employing a graphical user interface (GUI) for aeronautical information specialists (AIS) accessing NAS data. The primary requirements were that it be a reliable and maintainable system with simplified access to NAS data. The database includes over 1,500 data elements in over 350 tables, and provides access to the data by virtue of four temporally-different views; the analysts use over 400 user forms to access and manage NASR's data contents. Looking towards the future, interfaces with data sources and those systems dependent upon NASR information are being automated. It is planned that users will have access to an external website holding a copy of the latest database contents updated on a daily basis.

Many non-technical factors bear on the success of new transportation technology applications. Both the technology and transportation communities often underestimate the difficulty of achieving technological innovation in transportation. Crafting of wise and informed policies, plans and decisions in this complex arena requires a broad perspective and a clear vision of the complexity of the innovation process and the primary obstacles to its success. This paper suggests some key topics and issues that need to be considered when determining whether, when and how to direct major effort and investment toward application of advanced and emerging technologies in transportation, particularly in the context of innovative system concepts. In addition to uncertainties inherent to application of innovative technologies and concepts, a wide range of real-world challenges must be considered: broad understanding of user needs, behavior, and preferences; identification of implementation and transition difficulties; assessment of the capacity of new concepts for long-term evolution in technology and function; overcoming of political factors, institutional complexities and stakeholder conflicts; achieving compatibility and interconnection with existing technologies and systems; and adaptability to future changes in transportation needs, patterns and behavior. These topics are discussed primarily from the perspective of major public sector programs, within the political, financial and institutional environment of the US, but most of the issues considered are highly relevant in other contexts. A framework for assessing innovative approaches is suggested, illustrated by reference to the evolution of our existing transportation system, and by mixed results obtained in achieving technological innovation in recent decades. This discussion is relevant not only to policy on technology and infrastructure investment, but also to formulation of transportation R&D programs in these areas.

A limitation of most traditional methods of traffic data collection is that they rely on techniques that are strictly local in nature. Airborne sensors can provide sufficient resolution to sense vehicle locations and movements across broader spatial and temporal scales. Based on two recent "proof-of-concept" experiments, this paper describes the logistics involved in collecting and analyzing aerial video data, a general method for analysis of traffic data, specific methods for travel time and delay estimation, and an algorithm for automated vehicle tracking. Digital video, global positioning systems (GPS), and automated image processing are used to improve the accuracy and cost-effectiveness of the data collection and reduction.

Most Transportation planning, design, and improvement projects require the knowledge of the roadway geometry such as the banking or the cross slope. Although some of this information may be extracted from existing road plans, some existing roads were constructed without any formal design. Also, construction procedures may result in existing characteristics not matching the design characteristics. This paper discusses a methodology that uses a multi-antenna-single receiver configuration also known as attitude GPS to collect such data. Results from this study are presented together with system adequacy information and recommendation to the practical application of this system.

While there is general knowledge about how transportation affects the environment in urban and rural areas, remote sensing, image processing and GIS technologies provide new opportunities for incorporating environmental factors in infrastructure asset management, such as the extent of erosion, wetland displacement and destruction, runoff discharge, loss of tree cover, and residential and business dislocation. These technologies can be used to develop data for strategic monitoring and analysis of the impacts of various civil infrastructure development alternatives on the environment. This paper discusses the combined use of remote sensing, image processing and GIS techniques for infrastructure and environmental capital asset management. The authors discuss important considerations for developing RS/GIS databases and analytic approaches, to enhance existing decision support systems, and present examples of the types of analysis that could be performed to integrate transportation infrastructure and environmental asset management. Integrated infrastructure and environmental capital asset management will address the inevitable interactions that occur between the built, natural and social environments, and aim to use the acquired knowledge for identifying better alternatives for civil infrastructure development.

The Ground Penetrating Radar (GPR) based analysis of road pavement is actually performed by many agencies involved in road networks management. The most difficult step of this technique is the GPR image interpretation as previous discussions of the theoretical basis of this approach for the diagnosis of pavement damages with regard to the international catalogues of pavement damages and to the traditional techniques of monitoring has stood out. Software assisted processes have usually such a low reliability, that a road pavement engineer advisor is needed. The main intention behind this paper is to propose an optimal algorithm for GPR signal processing, in order to check rapidly the difference between a "linear continuous" and a "point located" singularity of the pavement. In the first case (linear) a layer thickness or geophysical characters modification is diagnosed, otherwise (point located) voids, water suction or cracks are diagnosed. The algorithm is based on the time delay computing of the GPR signal. This delay can be recursively expressed as τ_i^j = τ_i^j–1 + e where i is the space abscissa and j the time index, e is the error induced by a singularity, and, generally, the signal is y(t) = Σ_i=1,Na_i h(t–τ_i). A threshold analysis of the error is carried out to check any damage and, subsequently, if a damage is predicted, a second threshold analysis points its expected nature out. This approach has given interesting and promising outcomes. The difficulty to identify very thin layer (< 5cm) is also considered in order to diagnose damage in the upper pavement. An experimental survey has been performed either in laboratory and on road and the results show a good performance of the electronic system.

Many different methods of inventory data collection are used by transportation agencies in the United States, but are often time consuming and labor intensive. With data requirements for transportation agencies increasing, there is a need for more efficient data collection methods. This paper describes the application of remote sensing for inventorying transportation features. Remote sensing is the process of recording information about an area by measuring reflected or emitted radiation using a sensor that is usually based on an air or space platform. A pilot study was conducted to evaluate the use of remotely sensed images for the collection of roadway inventory features. Images individually collected at 2-inch, 6-inch, 24-inch, and 1-meter resolutions were evaluated and compared for the study area. Images were tested for positional accuracy, the number of inventory features that could consistently be identified, and ability to measure linear features.

A method based on Fisher Information from sensitivity matrix is presented for sensor placement in nondestructive tests used for parameter estimation and structural model updating using sparse and noisy measurements. This method can be used for both static and modal nondestructive tests but in this paper sensor placement for only the static nondestructive data is presented. An error function that represents the difference between analytical and measured data is used in this research. This method ranks the sensor locations and load cases using the sensitivity matrix that represents the rate of change of the error function used with respect to each unknown stiffness parameter. The proposed method is applied to a single story frame to demonstrate the selection of strategic sensor locations. This example shows that this is an effective and robust method for strategic sensor placement with regard to the stiffness parameters to be estimated.

Pavement management systems (PMS) help transportation agencies make pavement maintenance, repair, and rehabilitation decisions. Given the amount of data collected from pavements every year and the extent of information associated with managing pavements, one of the basic requirements of a PMS is a comprehensive database. Other components of the PMS include analysis routines, report generation modules, and a feedback mechanism. These components rely on the use of models to predict future network conditions, assess rehabilitation needs and costs, and select effective rehabilitation strategies. Deterioration models depend on the quality of data used for their development. Therefore, it is important that the database contain reliable, objective and appropriate information to support decision-making for planning and budgeting as well as to support the development of deterioration models. Ideally, a PMS database would contain enough data to satisfy all the requirements for developing such models; however, in practice, complete databases often are not available.

In this paper, we describe an automated method for detecting and correcting outliers in a data set generated by a weigh-in-motion scale. We use clustering and regression to identify outliers in an aggregated version of the data set. Then we compare the attribute distributions of outliers and the distributions of normal data to identify the mechanisms that cause the outliers. After we remediated these mechanisms, the outlier data closely resembled the normal data.

This paper demonstrates the usefulness of Falling Weight Deflectometer (FWD) testing for sectionalizing projects. Sectionalization involves dividing the project area (i.e., roadway, airfield, etc.) into sections, where each section has a fairly uniform pavement condition. This process facilitates identification of the weaker pavement sections of a project and thus helps to optimize rehabilitation efforts. In the project described herein, FWD testing was performed before and after a SUPERPAVE resurfacing program (involving 50 mm (2 in.) of milling and 100 mm (4 in.) of asphalt concrete resurfacing) to evaluate the effectiveness of the resurfacing effort. The project area consisted of a 12 mile segment of eastbound Interstate 78 in New Jersey. The FWD test results indicated that the median effective structural number of the pavement (SN_eff) for the entire project length had increased by 1.6 (resulting in an average remaining service life of 16 years) after milling and resurfacing of the pavement. These results, however, also indicated that a few sections of the roadway had higher deflections, lower SN_eff values, and a remaining service life as low as 8 years. These weaker sections were thus in need of a more extensive rehabilitation effort to provide a more uniform service life throughout the project length.

Inspection activities, such as bridge inspection, are still mostly done using a paper-based forms. However, with recent advances in computer technology, where the hardware is getting smaller in size, yet more powerful, computer systems are being employed and used in the field to help inspectors collect inspection information. In order to provide useful and practical computing support for inspectors in the field performing inspection activities, a computer system for supporting these inspectors has to be designed and customized to recognize the specific inspection task being performed and the context of the inspection. Unfortunately, the development of field inspection support systems is both difficult and time consuming since it suffers from a lack of generality, such that effort is wasted when implementing different applications sharing some common aspects. A framework that provides a partial solution for developing field inspection support systems was developed. This framework captures the common aspects of various field inspection activities so that the development effort of those field inspection support systems is reduced. This paper presents the application of this framework to the real world bridge inspection problems.

During construction, inspection, maintenance, and repair of transportation infrastructure projects, field engineers and technicians frequently refer to maps and other technical documents to check the location of structures and structural elements, and to collect data related to these structures. The concept of Mobile Augmented Reality System for Infrastructure Field Tasks (MARSIFT) will not only allow users to automatically retrieve the necessary information in real-time, based on the location and orientation of the user and within the specific task context, but it will also display this information as augmentation to the view of the surrounding real world objects such as roads, bridges, and tunnels. In this paper, the potential and limitations of the MARSIFT concept from the points of views of tracking, 3D GIS/CAD databases, and user interactions are discussed.

The Uinta National Forest is responsible for regulating the travel of vehicles in its forest to ensure motorist safety and preserve the protected environment. Road signs are the primary tools used by the Forest Service to meet these objectives. These signs, however, are subject to weathering and vandalism. The older sign management method consisted of paper inventory sheets that were unkempt, missing inventory sheets, and sometimes unreadable; it was difficult to effectively manage a vast number of signs. In order to expedite the process of sign management and planning a computerized sign management plan was developed and implemented to provide a framework for an effective sign management program by the Uinta National Forest Service. This system, named as the Sign Inventory Program was developed on computer spreadsheets that can be accessed by all pertinent Forest Service officers through the Forest's internal computer network. The inventory includes, among other information, the status of existing signs, a record of sign maintenance, photos, replacements, removals, and recommended new installations for traffic signs by engineering analysis. This computerized new system saves the Forest Service officers significant amount of time, compared with having to look through the entire inventory of signs and manually distinguish how many signs need to be replaced. A pilot program that integrates the Sign Inventory System with GIS systems and Power Point slide shows was also developed to provide a graphical illustration of the road along with links to inventory spreadsheets of each corresponding sign on the road. The Forest Service officers can search this GIS system by sign type and sign condition.

Traditionally, condition has been the desired metric for establishing maintenance and repair (M&R) needs for building components. Different approaches have been developed for that condition assessment, including the use of condition indexes. While having a condition metric is critical for identifying M&R work needs, this metric does not tell the complete story. Often, facility work needs go beyond degradation based maintenance, repair, rehabilitation, or replacement because components or systems also require capability upgrading or altering due to a change in user needs, technical obsolescence, and/or the invoking of regulatory requirements. These capability based needs can be significant and in some instances, exceed the condition based needs. Therefore, there is a need to have a metric for measuring capability that also can be used to support "network level" work planning. The U.S. Army Engineering and Research Development Center, Construction Engineering Research Laboratory (ERDC-CERL) has, in the past, developed inspection based condition indexes for building components, systems, and the overall building. ERDC-CERL has developed capability metrics for building components and systems. These capability metrics will supplement the previously developed condition indexes and together measure of "functionality" which is the total ability of a component to do what it is intended to. This paper discusses condition metrics as used in BUILDER EMS and the development of the capability metrics for the various building components. These metrics are intended to support "network" level management activities for communities ranging from relatively few buildings to large numbers of buildings such as found at large military installations and college campuses.

In the past three decades, infrastructure management has evolved from individual systems that manage one type of infrastructure asset to integrated asset management systems. This paper makes the case that systematic asset management is today's best approach for balancing growing demands, aging infrastructure, and constrained resources. Furthermore, it emphasizes that the existing process may be improved by taking advantage of novel soft computing technologies. Given the amount of uncertainty and imprecision encountered in the infrastructure inventory and condition data, the incorporation of soft computing technologies into the decision-making process of maintaining our transportation infrastructure systems may enhance tractability and robustness of the proposed solutions, and may allow a better rapport with reality.

Over the last two decades, awareness of the importance of maintaining the nation's investments in transportation infrastructure has increased. One manifestation of this concern was the management systems requirements in the Intermodal Surface Transportation Efficiency Act of 1991. While many of the decision-making and modeling techniques developed for these systems are reasonably mature, the extent to which the systems are actually integrated into the policy process is less than might be expected. A number of institutional barriers exist, but the reliance of existing systems on visual condition assessment data may also contribute. The last decade has also witnessed revolutionary developments in structural sensing and non-destructive evaluation technologies. In particular, the emergence of fiber optic technology and micro electromechanical systems (MEMS) has allowed high-performance interrogation of infrastructure components at data rates, channel counts, and resolution levels never before achievable. These sensing systems, when used in conjunction with specialized data mining and analysis algorithms, provide an integrated approach to in-situ infrastructure condition assessment. This is particularly important as owners seek to reduce costs associated with condition assessment, corrective action, and prolonging service life. However, with the exception of isolated demonstration projects, these state-of-the-art technologies have not been applied widely to condition assessment of transportation infrastructure. Although the techniques described can apply to pavements, geotechnical features, and other components of transportation infrastructure, this paper focuses on bridges. The paper briefly describes the state of the practice for bridge condition assessment. Next, the paper reviews fiber optics, as one of the promising state-of-the-art sensing technologies, and some key analysis approaches taken in analyzing data with the goal of condition assessment (at a point in time) and health monitoring (changes over time). We then discuss two applications of the sensing and analysis methodologies. The paper concludes with an appraisal of the future of these techniques in transportation infrastructure condition assessment.

Accurate data collection and interpretation of pavement data is critical for the decision-making process in pavement management. Collection of several data types for pavement management is automated and widely used. However, collection and analysis of pavement surface distress is still a manual process for many highway agencies, even though substantial amount of resources were used in the past decades to devise automated approaches for collecting and analyzing pavement surface distress. This paper introduces a new automated system capable of collecting and analyzing pavement surface distress, primarily cracks, in real-time through the use of high resolution digital camera, efficient image processing algorithms and multi-computer, and multi-CPU based parallel computing. The paper overviews the major steps in the algorithms for image processing. It is shown in the paper that distress results from the automated system are consistent for multiple passes of the same pavement sections.

Subgrade soil modulus is a required parameter for the AASHTO method of flexible pavement design and an indirect input parameter for rigid pavement design. Falling Weight Deflectometer (FWD) testing often performed to determine the insitu subgrade modulus of in-service pavement structures for pavement rehabilitation. Seasonal variations of subgrade modulus often complicate the pavement design process and hence the 1993 AASHTO flexible pavement design guide introduced an effective subgrade modulus concept to estimate the effect of seasonal variation of subgrade modulus on pavement performance. This study was conducted to determine the seasonal variation of subgrade modulus values with three types of subgrade soils; clayey sand, silty sand and poorly graded sand. Long Term Pavement Performance (LTPP) program under the Strategic Highway Research Program (SHRP) provides a substantial data source to evaluate the seasonal variation of subgrade modulus values. FWD deflection data were back calculated using DAPS back calculation program and evaluated with different subgrade soil types. This evaluation provides a better insight for future research to estimate an effective subgrade modulus ratio for different subgrade soil types.

A recursive nonlinear model was developed to predict pavement deterioration. Pavement deterioration was assessed in terms of loss of serviceability — expressed as a function of traffic characteristics, pavement structural properties and environmental conditions. The model highlights some of the advantages of relaxing the linear restriction usually placed on the model specification. First, a functional form that better represents the physical deterioration process can be used. Second, the estimated parameters are unbiased because of proper specification and the use of sound estimation techniques. Finally, the standard error of the regression was reduced by half that of the equivalent existing linear model. The model developed enables the determination of an unbiased exponent of the so-called power law, and equivalent loads for different axle configurations. The estimated exponent confirms the value of 4.2 traditionally used. However, equivalent loads estimated for different axle configurations differed from traditionally used values. The estimated equivalent load for a single axle with single wheels is 44 kN, while the equivalent load for a tandem axle with dual wheels is 148 kN.

To characterize in-service truck tire pressure and truck configurations on Texas highways, a survey on major Texas trucking routes was conducted. Analysis of the collected tire pressure data showed that factors such as axle weight, tire temperature, geographic area, highway class, traveled distance, and axle type, contributed to in-service tire pressure values. In this study, a factorial experiment design and stratified sampling method were employed. Analysis of variance, analysis of covariance, and t tests were used for the comparison of truck tire pressure data; laboratory experiments and linear regression were used to identify relationship between tire pressure and tire temperature. Comparisons and analyses of factors differentiating truck tire pressure are presented.

Pavement cracking models provide useful predictions of cracking initiation and progression, for purposes of pavement design and management. Pavement cracking initiation is a highly variable event. The widespread use of deterministic models of pavement cracking initiation is inherently problematic, as these models do not provide information on the variability of the cracking initiation times. Another common problem in modeling pavement cracking initiation is the inappropriate treatment of data censoring. Data collection surveys are usually of limited length. Thus, some pavement sections will have already cracked by the day the survey starts; others will crack during the survey period, while others will only do so after the survey is concluded. If the censoring of the cracking initiation times is not accounted for properly, the model may suffer from statistical biases. In this paper, an analysis of the pavement cracking data collected during the AASHO Road Test is presented. The analysis is based on the use of stochastic duration modeling techniques. Duration models enable the stochastic nature of pavement cracking initiation to be represented as well as censored data to be incorporated in the statistical estimation of the model parameters. The results presented in this paper show that the cracking initiation model provides good fit to the data, the parameter estimates have the correct signs and relative magnitudes, and the model predictions are more accurate than those obtained with the original AASHO model.

Sound infrastructure deterioration models are essential for accurately predicting future conditions which, in turn, are key inputs to effective maintenance and rehabilitation decision-making. The challenge central to developing accurate deterioration models is that condition is often measured on a discrete scale, such as inspectors' ratings. Furthermore, deterioration is a stochastic process that varies widely with several factors, many of which are generally not captured by available data. Therefore, probabilistic discrete state models are often used to characterize deterioration. Such models are based on transition probabilities which capture the nature of the evolution of condition states from one time point to the next. However, current methods for determining such probabilities suffer from several serious limitations. An alternative approach addressing these limitations is presented in this paper. A probabilistic model of the time spent in a state is derived and the approach used for estimating its parameters is described. Furthermore, a methodology for determining the corresponding state transition probabilities from the developed duration model is presented. Finally, the overall methodology is demonstrated using a data set of reinforced concrete bridge deck observations.

This paper presents a two-level approach for the maintenance management of aging highway bridges that integrates two different probabilistic deterioration prediction models. The level-1 management is based on Markovian cumulative damage models that predict the macro-response of bridge structures. The level-2 management is based on quantitative reliability-based mechanistic deterioration models that predict the micro-response of bridge structures. The level 1-management identifies the critically damaged structures and forecasts the overall deterioration and required maintenance funds for both short and long term planning for a bridge network or a bridge component. The level-2 management focuses on the critical structures that were identified from the level-1 management or from a specific condition assessment to evaluate their safety and serviceability, and optimize their maintenance. The proposed two-level approach to bridge maintenance management, which integrates two different probabilistic deterioration models, will help improve the effectiveness of maintenance management systems in satisfying the safety, serviceability, and budgetary requirements of highway agencies.

One of the most difficult tasks in pavement management information systems is establishing the links between performance measures of a structure and the design and construction inputs. During the last two decades Wisconsin Department of Transportation (WisDOT) has made significant steps in improving material properties, design practices, and monitoring of pavement performance. The current challenge is integrating these databases for effective pavement management, and performance-related design and construction specifications development. In order to fully understand the component of pavement performance and specify the inputs through specifications to achieve that performance we must develop quantitative relationship between input variables and response variables through a scientific, fully integrated system named pavement performance analysis system (PPAS). From such a perspective, this paper addresses the functional requirements of a PPAS and present the dimensional model in the context of data warehousing for developing a PPAS.

This paper presents the development of a pavement rutting model by combining experimental and field data. The experimental data came from the AASHO road test, while the field data was obtained from the Mn Roads Test site. Using joint estimation in this context allows us to improve the efficiency of the statistical estimates, quantify the values of the bias parameters in the experimental data, and identify some parameters that cannot be identified by using field data alone. The results of our research show that the pooling of experimental and field data yields better models than can be achieved by using a single data source.

The Illinois Department of Transportation (IDOT) currently uses Condition Rating Survey (visual rating) in managing its pavement network. Presently, decisions on when and which sections to rehabilitate are based heavily on CRS and the discretion and judgement of staff involved in prioritization. IDOT also collects pavement data, such as IRI, faulting and rutting for the highway network. However, they are not currently used consistently in prioritizing pavement rehabilitation. An effort was made to incorporate these additional factors in the decision making process. Different pavement ratings address unique conditions of a pavement section; CRS addresses the physical condition of the pavement (such as cracking), whereas IRI addresses the smoothness of the pavement, thereby considering user comfort. Moreover, rutting incorporates safety and traffic addresses usage of the facility. CRS and IRI can be utilized as separate pavement indices on which pavement prioritization can be based. However, a composite index of CRS, IRI, traffic and rutting would not only change the priority of pavement sections but would add additional credibility to the selection of projects for rehabilitation. An evaluation of Illinois Interstate highway network was conducted utilizing CRS, IRI separately and a composite index of CRS, IRI and rutting. The composite index is part of a weighted benefit ranking procedure which measures the overall benefit of improving (rehabilitation and reconstruction) pavement condition. This paper evaluates the consequences of utilizing the composite index versus CRS and IRI in prioritizing pavement rehabilitation on Illinois' Interstate highway network. The paper also discusses the weighted benefit ranking procedure and its implementation in a Geographic Information System (GIS) based pavement management system (PMS). It is expected that the consideration of this broader based composite index would produce a better selection of projects that would be more optimal to the highway user and highway agency.

Departures from design/construction quality requirements can surface at different stages of the project life. They include problems during the construction process, at the intermediate and end products, and during the operating stage of the project. This paper describes how problem patterns can be identified and analyzed for diagnosing and/or predicting nonconformity of constructed facilities. This will enable appropriate actions to be taken for eradicating the causes of nonconformity, and preventing their recurrence and/or their occurrence. Highway construction has been selected as an application and illustrative domain. Pattern analyses have been directed at deriving root cause classes of problems including: (1) Design, which relates to the assigned specifications, methods, and/or procedures; (2) Execution, which involves errors or the inability to execute tasks; and, (3) External, which includes unforeseen events or accidents. The proposed approach will provide construction personnel with an effective tool to analyze root causes of problems and to better reason about quality-related problems.

A decision support system has been developed to assist pavement management agencies in maintenance and rehabilitation (M&R) planning and implementation in the road network of Greece. Optimal strategy selection in the network level is obtained via a linear programming model which aims to minimize agency costs subject to constraints related to the desirable pavement condition over the network and planning horizon. In the project level, decisions about M&R project structure, planning and resource allocation are provided. A systematic process for data collection and analysis has been initiated and is expected to contribute to future system enhancements.

In this paper we discuss the development of a comprehensive pavement management system and its implementation in Portugal. The modeling approach is network optimization via a Markov decision model and includes several new features.

Under the traditional approach to develop maintenance and repair policies for infrastructure facilities, Policy Evaluation and Policy Selection are performed while assuming that a complete and correct facility deterioration model is available. In addition, the time, cost and complexity required to develop such models are ignored in the framework. In order to address these limitations we formulate the problem of developing maintenance and repair policies as a Reinforcement Learning Problem. Under this approach, it is not necessary to model a facility's deterioration process to perform Policy Evaluation and Policy Selection. These functions are accomplished through mapping the effect of actions prescribed by the policies. In this paper, we explain the agent-system interaction considered in Reinforcement Learning. We discuss the probing-optimizing dichotomy that exists in the process of performing Policy Evaluation and Policy Selection. Then we describe Reinforcement Learning methods that can be used to address the problem of developing maintenance and repair policies. Finally, we present the results of a simulation study where we show that Reinforcement Learning is a viable approach that can be used to adapt and fine-tune policies in situations where complete and correct models of facility deterioration are not (yet) available.

Physical infrastructure assets, such as roads, tend to represent major financial investments, be designed to last for many years, be heavily used, carry considerable load, and be exposed to the natural environment. Expenditure on constructing and maintaining such assets is considerable. For example, in Australia, it is estimated that about $A7 billion dollars are spent annually on roads. Allocation of the capital portion of such expenditure tends to be on perceived value for money, tempered with social considerations. The maintenance, or recurrent portion (which can be a quite high proportion of the whole) typically tends to be based on previous performance and "as seen" performance of the roads. While such a process is useful at a day-to-day operational level, a more strategic approach, in which the behaviour of the infrastructure in question and its likely future performance were better understood, would be better for making funding decisions and for focusing expenditure where it is most needed. As the various phases of the infrastructure life cycle interact, any strategic infrastructure management approach should consider the whole life cycle of the infrastructure rather than a particular segment of the life cycle in isolation from the rest of that life cycle. This paper considers the application of such a life cycle methodology to the infrastructure development, management and investment decision-making process.

In this paper, the authors continue the discussion of the TFC system evaluation framework and plan developed by Collura and Plotnikov (2001) and address the issue of TFC system performance evaluation measures. The authors suggest that cost-effectiveness and labor-intensiveness of TFC operations would be useful evaluation measures to assess the operating costs of existing and new TFC systems. The authors also present statistical data on selected heavy rail systems in the U.S. to illustrate the use of the proposed evaluation measures. In addition, these data support the notion that cost-effectiveness and labor-intensiveness may vary considerably among heavy rail TFC systems and that this variation is due, in part, to the TFC technology being used.

An intermodal control system (ICS) has been designed for the Long Island Railroad (LIRR) to minimize conflicts between trains and intersecting at-grade highways. A field test was performed in June 2001, at New Hyde Park (NHP), NY, a suburban passenger station with a nearby highway crossing. Because deployment will not occur for several years, a benefit-cost evaluation of this system is prospective, based on data obtained before and during the tests, and from simulation results. The project is of particular interests because it includes both the control of trains for the benefit of highway users and the control of highway users for the benefit of railroad operations. So far, the quantitative analysis presented here relies on illustrative trial values that are intended to be within the range of plausibility, but will require considerable further refinement.

Up-to-date congestion data are critical for evaluating improvement strategies that focus on mobility and system reliability such as corridor improvements, signal coordination efforts, and deployment of ITS components. This paper describes a tool for measuring travel time and delay along arterial corridors using global positioning system (GPS) and geographic information system (GIS) technologies. The tool is called GPS-Based Evaluation of Travel Time (GETT). GETT uses linearly referenced GPS data and checkpoints along the routes of interest to calculate partial and cumulative travel times, speeds, and delays. The application also measures three elements of intersection delay: control delay, approach delay, and stopped delay. The application produces detailed and summary tabular reports, and output data can be used to generate color-coded maps depicting congestion problems along corridors.

Real time kinematic (RTK) GPS is featured by its outstanding accuracy in measuring position and speed dynamically. A car-following experiment was conducted using ten passenger cars each equipped with RTK GPS receivers in a test track of Hokkaido Developing Bureau, Japan. This experiment has tested various speed patterns for lead vehicle including four sinusoidal curve patterns, four constant speed patterns and a random speed pattern. The effect of change in lead vehicles' speed on the driving characteristics following drivers were measured in terms of position and speed data taken by RTK GPS receiver equipped in each vehicles. These data are analyzed to establish the effect of change in speed in terms of variation of reaction time, sensitivity factor and stability factor along the platoon for each speed patterns. Although for some of the driving patterns, the variation of these characteristics of drivers seems to have some trends but it cannot be verified statistically as in most of cases no such trends are observed.

The focus of this paper is on using AVL (Automatic Vehicle Location) and APC (Automatic Passenger Counter) empirical data to develop a travel time model capable of providing real time information on bus arrival time to passengers (traveler information services) and to transit controllers for applying proactive control strategies. Three techniques have been used to develop bus travel time prediction models. The evaluation of these techniques showed that the time-lag recurrent neural network model outperformed other models in terms of accuracy and transparency demonstrating its ability to update the model based on new data that reflect the changing characteristics of the transit-operating environment.

Understanding car-following behavior is crucial in determining highway capacity, level of service, and safety. It is also the foundation for traffic simulation models that are commonly used today for traffic management. In the past, the car-following experiments were mostly conducted under controlled environment with only two vehicles. This research investigates the car-following behavior based on experiments consisting of three vehicles. The position and speed of each vehicle were recorded at one-second intervals by a high-power GPS receiver, based on which the distance headway between the vehicles as well as the acceleration rate of each vehicle were computed. Since the experiment was conducted on urban streets and freeways with the normal traffic, the data analysis has revealed the realistic car-following behavior that can help to better understand today's complex traffic flow characteristics.

Expenses associated with collecting data to produce estimates of annual average daily traffic (AADT) and vehicles miles traveled (VMT) could conceivably be reduced by using high-resolution satellite imagery. Vehicles can be detected in 1-m satellite-based panchromatic imagery now available to the civilian world. However, the short observation period and low sampling frequency of high-resolution satellite sensors pose a potential impediment to using satellite-based data to improve AADT and VMT estimation. We address this difficulty by developing computer software to simulate AADT and VMT estimation errors with and without the use of satellite data. Results indicate that adding satellite-based data to ground-based data would decrease AADT and VMT estimation errors and allow for a substantial reduction in ground-based samples for a large range of inputs. Other numerical results show that using satellite-based data could lead to improved estimates while reducing the number of the permanent automatic traffic recorders used to determine temporal adjustment factors in traffic monitoring programs and that there are decreasing marginal benefits from increased satellite supply.

Traffic management methods often employ network flow models that use data collected by a surveillance system to estimate the current state of the network and predict its evolution over time. Origin-Destination (OD) flows comprise one of the main elements of the network state. This paper focuses on the role of various types of surveillance data in the real-time estimation of dynamic OD flows. Traditionally, traffic surveillance is carried out using inductive loop detectors. While providing an inexpensive means of monitoring traffic, such detectors do not provide any data on traffic conditions beyond the detection point. Recent advances in communication, computing, and electronics are enabling the development of remote sensing-based surveillance systems which involve the observation of the roadway network from a fairly high vintage point. Such sensors can provide extended spatial coverage of the network, thus, capturing vehicle trajectories, queue lengths, and link travel times. Such data have the potential to improve the accuracy of the network state estimates along with traffic predictions and, consequently, result in more effective traffic management. This study quantifies the value of using intersection turning fraction and link travel time measurements, which are available from remote sensing-based surveillance systems, in estimating OD flows in real-time. A network consisting of three adjacent intersections is used to conduct an empirical analysis.

The TRansportation ANalysis SIMulation System (TRANSIMS) team at Los Alamos National Laboratory evaluated the proposed conversion of Diamond Drive on Laboratory property from four to three lanes. This conversion was meant to address safety issues caused by left turns from Diamond Drive onto two crossing streets (visibility problems for oncoming traffic) and the lane widths being non-standard (too narrow). Widening the road was not considered an option because of the expected high cost. For this analysis, the following situations were evaluated: 1) A 4-lane versus 3-lane road; 2) Morning and evening peak traffic; 3) Merging north versus merging south of the largest intersection for the proposed 3-lane case, morning peak; and 4) Potential traffic flux uncertainties of up to 20% in the morning peak traffic. The results were quite dependent on the simulated signal timings, and can be summarized as follows: 1) In terms of travel time through the system, the 3-lane option would work, on the average, about the same as the 4-lane situation; 2) A few travelers would experience a 25–50% increase in travel time over the average travel time of 3.5 – 4 minutes; 3) The 3-lane option will place the morning traffic at close to its maximum throughput. Additional traffic volumes would be expected to degrade the system; and 4) There is a distinct, dramatic advantage to merging south rather than north of the Diamond/Jemez intersection in the 3-lane option.

This paper describes the ridership estimation procedure for the proposed Beirut Suburban Mass Transit System over the 2000–2015 time frame, and its implications on the selection of appropriate mass transit technology. The primary tool used in the ridership estimation is the Emme/2 network modeling and simulation package. The modeling effort reflected a number of data collection and statistical analysis efforts. Assumptions about economic growth, travel demand, infrastructure development, and policies were adopted to predict passenger demand for the mass transit system under a variety of plausible scenarios.

This work aims at evaluating alternative methodologies for analyzing the impact of activity-balancing policies as tools for improving urban traffic conditions in the specific context of Brazilian cities, taking the City of São Paulo as a background case (a ten-million inhabitants city with an eighteen-million inhabitants metropolitan area). The US literature on subjects such as house-jobs balance and urban form and their effects on transport is especially relevant for gathering points of view. After reviewing the justification, prospects and constraints for activity-balancing policies, alternative approaches for evaluating the potential effects are compared based on four points: the main data needed, methods of study, analysis units and types of results. Model based approaches are investigated on grounds of data availability (including software implementation) and theoretical capabilities (weighting model complexity). A range of options covering conventional four-stage models and integrated land use/transport models are considered. Then, a framework for selecting the method of study is sketched and the specific case of a study for São Paulo City is screened. Despite its peculiarities, such as the availability of a large series of OD data from 4 decennial household surveys, this setting is thought to be representative of other large cities (in developed or developing countries). The final goal is the definition of a sound approach for the real case study to be undertaken shortly. Due to the lack of professional experience on real case studies on the subject, the careful methodological and theoretical analysis is judged to be of considerable value.

The primary objective of this study was to develop a comprehensive dynamic model of household automobile transactions and choices at a disaggregate level to be used in a dynamic microsimulation modeling framework that can provide a direct forecast of consumer demand for personal-use vehicles given the available choices. A market-based decision making process and a transaction approach were applied for this project due to their consistency with the actual processes followed by a decision maker in a real world. The proposed framework has a nested tree with transaction choices (adding new vehicle to fleet, disposing one vehicle, trading one of the vehicles in fleet, or do-nothing decision) in the upper level and class and vintage choices in the lower levels. Different modeling approaches to operationalize such a framework were employed including hedonic price, multinomial logit, nested logit, artificial neural networks, and random parameter logit models.

Over the past five years, the threat of hurricanes in coastal regions of the United States (US) has resulted in several mass scale evacuations. Two of the largest evacuations in the history of the United States accompanied Hurricane Georges in 1998 and Hurricane Floyd in 1999. To better deal with future evacuations, emergency management and transportation officials in hurricane threatened states are seeking to apply traffic and weather sensing technologies to gain access to more timely and accurate information on the status of evacuation routes. Armed with this information these officials could more easily determine which routes are flowing smoothly and which are in gridlock so that traffic can be strategically redirected to nearby parallel routes for shorter clearance times, and used to optimize evacuation initiation and cut-off. Recently, the Louisiana Department of Transportation and Development (DOTD) and Louisiana State University (LSU) initiated a research effort to develop innovative and cost effective ways to gain access to evacuation route information. What has evolved is a partnership between the DOTD, LSU, the United States Geological Survey (USGS) and the private sector that seeks to create a state-wide traffic, weather, flood, and bridge scour monitoring system for critical routes in Louisiana. This paper summarizes the concept for the system and highlights many of the features and capabilities that will be incorporated into the system. It also presents some of the unique aspects of the system and the unique partnership that has been formed between the DOTD, LSU and USGS to make this system a reality.

A model has been developed specifically to target evacuation planning and operation around nuclear power plants. The software package is entitled TEDSS 3.0 "Transportation Evacuation Decision Support System". The model subdivides the 10-mile evacuation area, measured from the nuclear power plant as a center, into Emergency Planning Zones (EPZ). Evacuees are loaded onto the network based on the type of evacuation carried, the weather condition, and night or daytime, and slowly and quickly escalating evacuation. The user can modify through a computer screen interface the transportation network to satisfy the evacuation conditions and has the option to change the major routes from two-way flow to one-way flow. An internal module subdivides the network into best evacuation area for each EPZ centroid and consequently determines the best evacuation routes within that area. The outputs show the evacuation times, the evacuation routes, and the expected bottlenecks in the network. The model has been applied to develop the evacuation plans around Surry power stations in Virginia.

It is actually well known that road safety is crucially related to human factors. The road geometry itself often induces incorrect driving behaviors. This is confirmed by recent on road experiments, using an instrumented vehicle, such as by some occasional studies based on simulation. A validated driving simulation based approach is efficient and user friendly. This paper shows at first the outcomes of a laboratory survey that involved 120 drivers (> 2400 simulations) using a simple PC image generator. One entry straight, a transition curve (spiral) and a circular curve, a symmetrical spiral and a final straight geometrically compose the considered road. The dispersion of the trajectories for different speeds and lengths of the spirals and different radius of the circular curve has been analyzed. The outcomes show that usually assumed approaches in road geometry design can sometimes induce a deficit of safety. For example, when the angular deviation of the circular curve is greater than 40°, the safety standard decreases if the spiral length increases. This result seems to disappoint the well-established know-how. It cannot be explained through physical relationships, because of it refers to human perception and reaction. Secondly the paper presents more advanced results, that have been obtained using an interactive driving simulator. A road, in closed path configuration, is designed for simulation. A complete simulation survey has been carried out, recording the trajectories, the speeds, the accelerations, some bio-medical parameters, and more than 30 other parameters. This experimental survey greatly confirms the effectiveness of the simulation approach to validate the standards of the road geometry, for the safety assessment. Finally the study, sometimes contrary to the usual traditionally adopted technical rules, demonstrates it is not possible to set a geometrical safe standard, unless the impact on the human factors is verified.

Standards' validation of a driving simulation is needed to consistently approach the questions of road engineering and the impacts on safety. This simulation approach is used in order to take into consideration human factors in road accident generation. Road safety is not only related to physical and cinematic events but also to psychological factors, though the technical rules normally neglect them. Many psychologically based studies have shown how some internal and external stresses can produce unsafe actions of the driver. Sometimes the road itself directly induces these stresses. Recent on road experiments, using an instrumented vehicle, confirmed this theoretical framework, such as some occasional studies based on the simulation. On road surveys are usually very difficult and expensive to perform. Moreover it is possible to analyze only the existing road geometries and environments. Finally the collected data are not always stochastically significant. To overcome these difficulties a driving simulation based analysis of the safety of road is here proposed. The driving behaviour in simulation has already been validated. The aim of this paper is to identify the configurations of the virtual reality (VR) system relating to each different objective of the study. The dynamic equations and the electronic and physics characteristics of the interfaces are here considered. The parameters of the dynamic model are classified in road ones, such as geometry and pavement, vehicle ones, such as engine, shape/mass, and tire and environment ones, such as meteorological data. The interfaces are related to the visual, acoustic and mechanical feedbacks. This study numerically checks different configurations of the system through more than 10³ simulations, in order to identify the minimum standards in function of the different road engineering and safety needs. Previous research results in psychological and medical fields are here assumed and briefly verified, in order to assess the image generation rate, the auditory virtual environment, and each sensorial feedback. The considered configurations for the system have been performed and validated using the simulator of CRISS (STI driving simulator). The paper shows very promising frontiers and the implications and opportunities are discussed.

This paper presents a statistical description of light vehicle (passenger vehicles, sports utility vehicles, vans, and pickup trucks) crashes in terms of their major crash types, physical setting, and concomitant pre-crash scenarios based on the 2000 General Estimates System national crash database. This database contains variables that categorize the pre-crash situation and identify the vehicle movements and critical events immediately prior to the collision. This crash analysis supports the U.S. Department of Transportation's Intelligent Vehicle Initiative to solve traffic safety problems through the development and deployment of vehicle-based crash avoidance systems using advanced technologies. In the year 2000, light vehicles were involved in about 6,133,000 or 96.0% of all police-reported crashes on U.S. roadways. A breakdown of these crashes revealed 10 different crash types. Rear-end, crossing paths, off-roadway, and lane change crash types comprised the majority of these crashes with a combined frequency of 5,240,000 crashes. About 39.6% of all light vehicle crashes occurred at non-junctions, while 24.5% and 20.4% of all light vehicle crashes happened respectively at intersections and near intersections. Further breakdown of the four most dominant crash types identified 25 dynamically distinct pre-crash scenarios that influence the design of the sensory and decision-making elements for potential countermeasure systems.

Comprised of a broad range of diverse technologies, including information processing, communications, control, and electronics, Intelligent Transportation Systems (ITS) were first introduced as a means to solve the conflict between constantly increasing travel demand and insufficient transportation infrastructure. However, the impact of ITS on road safety was soon realized, and it became one of the major purposes of some ITS systems instead of only a side product. This paper summarizes the impacts of ITS technologies on road safety and the challenges of ITS safety evaluation. Based on the evaluation of collected ITS benefits information, this paper concludes that a variety of ITS technologies have the potential to make a substantial contribution to road safety and thereby have a substantial positive economic impact in the United States.

Cal Poly State University is completing several projects funded by the Caltrans New Technology and Research Program intended to develop, demonstrate, and evaluate an open architecture, low cost, advanced public transportation system (ARTS) suitable for small and medium size transit properties. From 2000 through early 2002, APTS technologies known as the Cal Poly Smart Transit System were implemented on San Luis Obispo Transit (SLO Transit) buses. This system provides: 1) GPS-based location of buses; 2) Emergency silent alarms for use by drivers; 3) Real-time radio transmission of bus locations; and 4) Solar powered Smart Transit Signs at key bus stops at Cal Poly and in San Luis Obispo to notify riders about the status of arriving buses. Most of this equipment was fabricated from relatively low cost, off-the-shelf components. This paper describes the APTS technology demonstration on SLO Transit and some findings of a parallel evaluation of the impacts of the new technologies on system performance and user satisfaction. The evaluation includes an analysis of bus operations prior to and after deployment of the new technologies, and surveys with riders. Although the technology is still new and the settling-in process is incomplete, findings to date are encouraging regarding the benefits of deploying appropriately scaled APTS technologies in such semi-rural settings.

Current transit network design methods often are over-simplified in terms of demand interactions, or neglect the importance of designing an integrated multi-modal network. In this study, we develop a two-phase methodology for integrated transit network design in a multi-modal network. Estimation of refined transit demand between origins, destinations and inter-modal transfer locations is processed in Phase 1; while the transit routes are built in the Phase 2. A numerical example is used to illustrate the approach.

This paper describes new adaptive fuzzy algorithms for coordinated ramp metering. The new model family named ACCEZZ (Adaptive and Coordinated Control of Entrance Ramps with Fuzzy Logic) was developed to overcome the limitations of existing coordinated ramp metering algorithms. The genetic fuzzy ACCEZZ models are explained, evaluated via simulation, and compared to other ramp metering approaches in a simulation scenario for a German Autobahn. One of the genetic fuzzy ACCEZZ models is currently installed with a comprehensive video detection system at the Olympic interchange of the Middle Ring Road in Munich within the research and development project MOBINET. The fast preliminary evaluation results of this real world implementation are very promising and a detailed numerical evaluation process is in progress.

This paper introduces GAPSIM, a hybrid Genetic Algorithm-Parallel SIMulation model for scheduling of lane closure requests that aims to minimize total traffic delay in a network over a 24-hour period. Genetic algorithm is used as the search engine while a microscopic traffic simulation tool is used to estimate traffic delay under a set of lane closure schedule. To save computing time, traffic simulations are run in parallel in the different processors of a multi-processor machine. To speed up the convergence to the recommended lane closure schedule, three improvement techniques, namely precondition, standard error criterion, and termination criterion, are also implemented. GAPSIM has been tested in an example problem involving 20 lane closure requests in a day in a suburban road network consisting of 986 links, 397 nodes and 22 origin-destination zones. The results show that, the parallel simulations, combined with the improvement techniques, produces 89% saving in computing time.

This paper presents two different Genetic Algorithms (GA) applied to design signal coordination for oversaturated networks. Signal coordination is formulated as a dynamic optimization problem and is solved using GA for the entire duration of congestion. This paper considers a tradeoff between simple GA (SGA), which requires a large population converging in a single convergence epoch, and micro GA (MGA), which requires smaller population with multiple epochs. A comparison is made for given resources available, that is, a fixed number of function evaluations. To provide quality solutions, SGA requires a large population but takes a longer time to converge, and thus it is not efficient for a real-time system. MGA overcomes the drawback encountered by SGA, that is, the time penalty involved in evaluating the fitness values for a large population. This paper reveals that MGA implementation on signal coordination problems reaches the near-optimal region of signal timing much earlier than SGA implementation. For a given number of functional evaluations, a small population size of MGA outperforms SGA executed with a larger population size.

This paper presents the comparison of Genetic Algorithms (GA) application to signal coordination for congested networks. Signal coordination is formulated as a dynamic optimization problem in which green times are decision variables and are represented in the individual GA candidate solutions. Two different types of serial GAs will be used depending on the way a set of candidate solutions is derived. The first is the Simple Genetic Algorithm (SGA). This GA uses three standard genetic operators: selection, crossover, and mutation, to generate a new set of solutions from the previous one. The second is the Bayesian Optimization Algorithm (BOA), which generates new candidate solutions using an estimate of the joint distribution of current promising solutions. The joint distribution is constructed using Bayesian networks. Investigation of each type of GA was made in terms of the number of functional evaluations needed to achieve predefined convergent criteria when the size of GA's candidate solution (population) is increased. Solution qualities are also compared. The growth of functional evaluations for SGA is close to a quadratic function with respect to the size of population, while that for BOA is a cubical function. For a small population size, SGA provides better quality solutions, while for a larger population size, BOA yields better results.

This paper describes research into integrating traffic micro-simulation models with driving simulators. It is hypothesized that the integration will improve the representation of individual vehicles in traffic simulation models by giving control of an individual vehicle to a test driver in the driving simulator. At the same time, the ambient traffic in the driving simulator that is encountered by the test driver will be improved as the traffic micro-simulation model controls the movement of each vehicle. The integration methodology is described and the potential benefits of the integration are discussed within the context of research on the passing maneuver.

This paper briefly presents the development of a high fidelity multilane traffic simulation model CELLSIM. The model formulation uses concepts of cellular automata (CA) and car-following (CF) models. The model is more sophisticated than CA models and simpler than CF models. The simplicity of modeling allows higher number of vehicles to be simulated. The model represents realistic behavior of traffic in congested and non-congested conditions. Development of straightforward algorithms for lane-changing are presented in detail in the paper. Lane changing is generated when a vehicle is forced to follow a slow vehicle, need to merge or diverge, or due to lane dropping. The lane changing logic replicates realistic behavior of drivers by utilizing space gaps and difference in speeds of leader and follower. The model is being validated macroscopically using three sets of field data. The model can be used to study real time traffic control for bottleneck and incident conditions. Moreover, the model has potential for managing traffic in a statewide traffic management system for disaster planning and evacuation.

The need to simulate a number of advanced ITS concepts prior to deployment necessitated development of high performance microscopic simulators for heuristic dynamic traffic assignment, freeway corridor diversion, driver information systems including variable message signs and vehicle guidance systems, real-time adaptive traffic control and other traffic management concepts. In spite of recent progress in developing sophisticated simulators for ITS applications their employment in real life projects requires substantial data collection, testing, validation and calibration. The latter is often ignored in practice where for convenience simulators are often used without sufficient validation and calibration, resulting in unreliable and/or inaccurate results. The objective of this paper is to demonstrate the issues and lessons learned in the employment of a sophisticated simulator for evaluating the effectiveness of adaptive ramp control in the Twin cities of Minneapolis and St. Paul, which is one of the earliest and most extensive worldwide (210 miles of freeway, 430 ramp meters).

Analysis of traffic conditions on freeways and arterial streets is critical to predict network performance. Aside from this, the results from the analyses are key inputs into funding decisions for construction, operation and maintenance of various projects. Several computer software programs exist that are capable of performing such traffic analyses and simulations. Software, such as TSIS/CORSIM, VISSIM, ITRAF and INTEGRATION to name a few, all claim to make accurate estimates of level-of-service, speeds, density, vehicle delays, emissions and other performance measures. However, each of these programs has their strengths and weaknesses. Their performance in terms of reproducing observed field conditions depend on the traffic conditions at hand and how well the models are calibrated. The main objective of this paper is to compare and evaluate CORSIM and VISSIM simulation software using default parameters for traffic conditions on basic freeway segments based on their ability to accurately reflect real-time observed field traffic conditions. Data were collected on a 2.5-mile basic freeway segment on US-95 in the Las Vegas metropolitan area. The observed speeds were in the range of 42 mph to 60 mph. Results from CORSIM and VISSIM show that the average speeds of operation are in the range of 55 mph to 60 mph. Both the software overestimated average speeds and underestimated traffic volumes when traffic volumes were high. However, both the software generated better average speeds when traffic volumes were low, though CORSIM overestimated traffic volumes significantly. In general, when compared with the real world data, in most cases CORSIM results are closer to reality than those obtained using VISSIM. However, there are many more such aspects, which need to be considered when selecting simulation software for analysis. For example, in CORSIM, speed generally means the average speed of a complete link considering all vehicles. However, average speeds can be calculated for each individual vehicle or for any given class of vehicles using VISSIM. The advantages and disadvantages of using the traffic simulation software to reproduce the real world data are discussed.

A key component in measuring the performance of a transportation network using simulation software involves simulating incidents and studying their impacts on system performance. Incidents on a freeway influence traffic behavior on not only the lane on which it occurred but also show considerable effect on the traffic of the adjacent lanes and the secondary road network. The capabilities afforded by various traffic simulation software to create and analyze incident vary significantly. The objective of this paper is to explore the capabilities of the CORSIM and VISSIM traffic simulation software to simulate and analyze incidents on a basic freeway segment. The study showed that incidents could be simulated more effectively using CORSIM when compared to VISSIM. The advantage in CORSIM lies with the comprehensive freeway incident simulation procedure, which is available as a part of the FRESIM module. The user can control the incident by specifying various factors such as the longitudinal position of a freeway link at which the incident has occurred, distance over which the effects last and the duration of the incident that influence the incident. The percentage of the traffic affected by the incident on the lane adjacent to the incident lane can be pre-established. This is used to reproduce real time situations where the traffic is affected by the adjacent lane traffic behavior. VISSIM has no special provision to simulate freeway incidents and/or work-zones that close a lane. However, incidents can be created on a freeway lane and its effects studied by exploring the capabilities of VISSIM by simulating a bus to stop for a stipulated amount of time. But, VISSIM does not have the ability to specify blockages or "rubber necking" on a lane specific basis or to simulate short-term and long-term interruptions to traffic (known as events) as can be done using CORSIM software. Data from the Las Vegas metropolitan area are used to compare the capabilities and performance of CORSIM and VISSIM simulation software.

Toll Network Capacity Calculator, TNCC, determines the optimum number of vehicles and/or passenger car equivalents that can pass through all available lanes at a toll collection facility in one hour. The capacity depends on the number of lanes available at the plaza and the processing rates for each of the customer-groups served in the lanes. The capacity also depends on the distribution of the arrival volumes among the customer-groups. In other words, the percentage of arrival volumes belonging to each of the customer-groups is an important factor in the determination of the maximum throughput. TNCC was applied to 58 toll plazas on the toll network in Orange County, Florida, maintained by the Orlando Orange County Expressway Authority, OOCEA. Four categories of vehicles or customer-groups were identified in this study. Videotapes of the throughputs at the plazas provided the processing rates for the customer-groups. The distribution of the arrival volumes among the customer-groups was extracted from transaction data. Once capacities were calculated throughout the toll network, potential bottlenecks were identified, generally centered around urban highway segments. At plaza bottlenecks, lane configurations were altered by TNCC and new capacities were determined. Through the software, it was possible to find the optimal lane configuration for maximum throughput at each of the 58 toll plazas.

This paper presents a status report concerning on-going research work by a team of Canadian researchers to develop a microsimulation-based, integrated model of urban land use and transportation. The paper briefly discusses the microsimulation approach to modeling dynamic, complex systems, in which the behavior of individual actors/agents is simulated over time and space. It then describes in some detail the overall design and current status of the ILUTE (Integrated Land Use, Transportation, Environment) modeling system under development. The overall purpose of ILUTE is to simulate the evolution of an entire urban region over an extended period of time. Such a model is intended to replace conventional, aggregate, static models for the analysis of a broad range of transportation, housing and other urban policies.

Traffic simulation, interfaced with traffic signal controller or traffic control algorithms, can be used to evaluate traffic operations and traffic control in a laboratory environment. This virtual environment can be integrated through client/server computing approaches. The CORSIM (CORridor SIMulation) is microscopic traffic simulation software developed by the Federal Highway Administration. It is distributed as a part of the Traffic Software Integrated System (TSIS). The TSIS suite is a collection of software tools used to support CORSIM. TSIS provides an Application Programming Interface (API) between an external application (e.g. a traffic control application) and CORSIM. This type of interface is known as a run time extension (RTE). To overcome the limitations of CORSIM RTE, a new type of CORSIM client/server API was proposed. The client API was defined and developed to interface CORSIM with a third-party virtual traffic controller server. The new client application needs few source code changes for different servers by different vendors. A railroad preemption case was designed to test the architecture and the client side software. Sample cases ran successfully in real time on a TCP/IP network with no program execution delays. This paper exemplifies how the client/server application can benefit highway railroad traffic signal preemption studies. The example cases show the dynamics of area traffic (including railroad traffic), traffic control devices and other factors have been systematically and realistically captured in evaluating railroad preemption design.

Given that it is extremely unlikely that the coming years will witness major capacity-expansion projects, transportation planners will now need to view the existing infrastructure as fixed, and start thinking about how much development the current system can sustain. This line of thinking, which involves deriving land-use limits from infrastructure capacity, requires solving the inverse of the typical transportation-planning problem. While there have been some reported studies in the literature that examined how to reverse the direction of the transportation planning process, the errors of the developed models were rather significant. In the current paper, we describe our efforts toward building on these earlier attempts in an effort to develop refined tools for solving the inverse of the transportation-planning problem. Specifically, our efforts have focused on developing artificial neural network (ANN) models for determining zonal trip ends from link volumes. The ANN models are developed and tested using real-world data from the Chittenden County region in Northwestern Vermont. Two different types of ANN models are developed, and their performance is compared. The study illustrates that ANNs are quite capable of solving the inverse of the transportation-planning problem.

The ability to accurately forecast train arrival times is essential for the safe and efficient operation of highway-railroad grade crossings (HRGC). Trains in the U.S. are required to give a minimum of twenty seconds of warning time before arriving at a HRGC. With the recent development of new detection equipment technology. detectors could potentially be employed further upstream of the HRGC which would result in earlier detection times. This information would be particularly useful for preemption strategies at signalized intersections located near the HRGC (IHRGC). For example, earlier warning times could be used to reduce or eliminate the risk of pedestrian movements at IHRGC being truncated in an unsafe manner. In this study, a modular Artificial Neural Network is used to forecast the train arrival time at a HRGC. An ANN was adopted because there is a non-linear relationship between the independent variables (i.e. train speed profile) and the dependent variable (i.e. arrival time at HRGC). A modular approach was used because the trains often have different characteristics depending on their cargo and the operational rules in effect at the time they are detected. The test bed was a railway corridor located in College Station, Texas. Doppler radar equipment was utilized to measure train speed and direction approximately 2.2 km from the HRGC. The arrival time forecast was performed 100 seconds after the train was first detected. Approximately 190 trains were used for training the ANN and 76 trains were used for testing. It was found that a modular architecture (with 4 categories) gave superior results to that of a simple ANN model, standard regression techniques, and simple forecasting methods. For example, the average absolute error was reduced by 29.6 percent as compared to the current method.

The model for real-time traffic flow prediction of intersections based on wavelet neural network is proposed. The main characteristics of this model are that the laws of traffic flow can be learned dynamically using wavelet network. Therefore, there is no complicated computation for the model learning process and it can also learn from historical data constantly. The experimental results show the effectiveness of the model with such properties as simple structure of network and fast convergence.

This paper investigates the use of neuro-fuzzy models for behavioral mode choice modeling. The concept of neuro-fuzzy models has emerged in recent years as researchers have tried to combine the transparent, linguistic representation of a fuzzy system with the learning ability of artificial neural networks (ANNs). Several neuro-fuzzy systems have been reported in the literature. They include various representations and architectures and therefore are suitable for different applications. In this paper, the performance of two of the most widely used neuro-fuzzy models, namely: B-spline Associative Memory Networks (AMNs) and Adaptive Network Based Fuzzy Inference Systems (ANFIS), is compared. The theoretical backgrounds of both systems are presented and their relative advantages are discussed using a mode choice modeling case study. Areas of comparison include: model performance, dealing with the curse of dimensionality, Automatic exclusion of irrelevant inputs, and model transparency.

Previous studies have shown that the current practice of highway geometric design, which is based on two-dimensional (2D) analyses, is a primary source of inaccurate driver's perception. Design guidelines, therefore, recommend designer to examine and adjust the alignment during the different design stages based on its three-dimensional (3D) visual quality. Tools that can be used in such a task include continuous roll of plan-profile papers, perspective drawings, and computer software. However, an established procedure with clearer and more quantitative guidelines would undoubtedly ensure less subjectivity and greater uniformity in alignment design. Such a procedure would require the quantification of several items pertaining to driver perception and behavior. This paper presents a summary of three different experiments carried out by the authors where computer animations were used to examine and quantify concerns related to driver perception. The objective of the first experiment was to determine the distance required by drivers to perceive the existence of a horizontal curve, which may be restricted by an existing crest vertical curve. The second experiment was designed to examine the hypothesis that horizontal curves appear sharper or flatter when overlapping with crest or sag vertical curves, respectively. Finally, the objective of the third experiment was to quantify the relationship between the actual and perceived radii. While all experiments utilized computer simulation of the road alignment, each experiment was designed differently to address its respective objective.

Traffic parameters are critical for traffic management system and traffic control. This paper is focused on developing a flexible and reliable system to detect the traffic parameters through image sequences. Image data are converted to digital string signals by virtual-loops. Through detecting and tracking the digital strings, vehicle's spatial-temporal features are extracted. In the end, multiple criteria and multistage decision are used to report the parameters of traffic load, traffic velocity and vehicle classification. The real-time prototype system is implemented on a computer. Experimental results illustrate the detection accuracy.

Recently, Lo developed a cell-based dynamic signal control formulation designated as Dynamic Intersection Signal Control Optimization (DISCO). DISCO considers the entire range of the Fundamental Diagram by encapsulating the Cell-Transmission Model (CTM). To determine an optimal timing plan, one may either solve the mixed-integer program directly or develop a quasi-optimal solution approach based on Genetic Algorithm. Lo et al shows that DISCO outperformed TRANSYT by more than 30% in terms of overall delay reduction when applied to Hong Kong network scenarios. However, both approaches require extensive computational effort. To rectify this shortcoming, this study develops a derivative-based method, combined with Frank-Wolfe's convex combination algorithm to determine an optimal timing plan. The derivatives are derived using Newell's three-point detector approach. A simple example is included to illustrate the approach.

This paper proposes a bi-level programming approach for the calibration of the parameters of a micro-simulation model where the origin-destination matrix (OD) is treated as a parameter. In the first level a genetic algorithm is used to calculate the micro-simulation parameter set given an OD. In the second level the OD is calibrated given the calibrated parameter set from level one. The CORSIM model was chosen for testing because the input data of parameter set and OD information can be easily updated during the calibration process. The proposed approach was illustrated on a section of I-10 in Houston, Texas and it was found that this approach resulted in a mean absolute error that was approximately 70 percent better than a standard, one level calibration technique.

Travel over intermediate distances is awkward. A typical air trip has a 600 mph line-haul speed with a 45 minute "access time" on each end. Cars offer much better access time but the line-haul speed is only 60 mph. This ten to one "speed gap" results in inefficient use of scarce infrastructure capacity, on both highways(everyone driving his or her own car) and runways (short-haul airline flights). There is a national need for a system with higher speeds than an automobile and lower access time than an airplane. Research in the area of high speed ground transportation (HSGT) has proceeded very slowly, but recently there have been some interesting developments. Most activity has been confined magnetically levitated vehicles (maglev). The Germans, who spent three decades developing their Transrapid maglev system, have abandoned their commitment to build a maglev system between Hamburg and Berlin. The Japanese maglev efforts appear to be having great difficulty making the transition from a research program to an actual system. In contrast, the Chinese have decided to jump into the maglev arena with both feet. First, they decided to buy the German technology and install a Transrapid system between Shanghai and the Shanghai Airport. Second, they are considering the construction of a repulsive maglev system using superconducting magnets in the city of Chengdu. In the U. S., maglev has been an on and off affair. As with Japan and Germany, there is resistance to making a commitment to actually building a system. The U. S. seems to be stuck in a dilemma — there is recognition that simply building more highways is not a wise choice, but there is also very strong preference for incremental investments rather than the type of lump sum financial commitment which has characterized previous HSGT concepts. The best way out of this dilemma is to identify a way that HSGT could be developed incrementally.

In this paper, a real-time bus travel-time prediction model is established by using Kalman filtering theory. The novel model can predict the travel-time by utilizing detected traffic volume on the links in the urban road networks, which is significant to the operation of urban buses system. The model is tested with practical detected data, and the results show that it has good properties.

This paper presents a quickest path tree algorithm designed for real-time use with urban road networks. Using piecewise linear functions to model arc times as a function of time-of-day, the algorithm finds a quickest path and its associated earliest arrival time from the origin node to every other node in the network, and for every desired departure time from the origin. Using sensitivity analysis, it transforms the min path tree for one departure-time interval into another for the next adjacent interval, whose width the algorithm determines on the fly. By building trees for only relatively few departure times, it determines quickest paths for every departure time. Preliminary tests show the algorithm running upwards of ten times faster than the conventional approach.

Traditional approaches to identifying and analyzing hazardous roadway locations, such as those presented in Missouri's Manual on the Identification, Analysis and Correction of High Accident Locations, consist of using statistics to generate a list of high accident locations (which are not necessarily hazardous), and to investigate and analyze each of the sites listed. There exist several problems associated with this approach which are discussed in detail in Zegeer and Spring. A 1993 study, by Spring, addressed problems associated with the exclusive use of statistics for identifying hazardous locations and resulted in the development of a knowledge-based geographic information system (KBGIS) whose prime focus was the identification of hazardous highway locations using engineering knowledge. Another complementary 1988 study by Spring resulted in the development of a demonstration prototype expert system (Hazardous Location Analyst-HLA) which analyzes hazardous intersections and establishes countermeasures for each. Subsequent studies subjected the system to tests, the results of which may be used to enrich and refine its knowledge base. The objective of this paper is to describe work completed at the University of Missouri-Rolla which examined the feasibility of a "turn-key" system for the site identification and analysis processes within the framework of the generic highway safety improvement program using a KBGIS approach.

Transportation agencies interact and exchange data with companies that operate utilities within the highway right-of-way (ROW). Unfortunately, obtaining up-to-date, compatible geo-referenced information from utility companies is frequently challenging. This makes the process of incorporating utility data into the transportation agency's own data management systems difficult. This paper describes a prototype utility data model that documents utilities located within the highway ROW. The paper also describes the architecture of a web mapping tool that allows existing and proposed utility installations to be displayed on widely used browsers. The tool can be implemented as a standalone application or as part of typical utility-related business processes at the transportation agency. In the current version, the web mapping tool supports online utility permit applications.

This paper proposes a new approach for modeling transportation systems. The proposed multi-layered modeling approach involves explicitly modeling information flow and the interactions between the physical and information flows. To illustrate this idea a simulation model of a runway is built using colored Petri nets. The model consists of two interacting layers: physical and information layers. The physical layer models the aircraft movements while the information layer deals with control and management of the air traffic. Aircraft communicate with the information layer to get clearance for takeoffs and landings. Communication links connect the two layers and carry clearance messages as well as other pertinent information. To investigate the impact of disruptions in the flow of information on system performance a number of scenarios are considered. Deterioration in system performance is measured in terms of average and maximum delays to the aircraft. Simulation reveals where the system is more sensitive to these information flow delays.

A persistent problem in household travel surveys (HTS) is the inability of subjects to provide accurate geographic information about their origins and destinations, and accurate, objective data about trip times and distances. Most people know their home address. Many workers know their workplace address; few parents or children know their school addresses. Few, if any, can provide the street address of the places where they shop, and many will not know the street address of friends and relatives that they visit in the local area. Other addresses, e.g., for churches, restaurants, banks, are not known. This results in lengthy and costly geocoding that still leaves 10 to 20 percent of trip ends uncoded, with a consequent loss of the data. Other addresses may be coded imprecisely, because it cannot be determined on which side of a street or intersection the visited location is sited. People report perceived travel times and distances. The content of the time, e.g., travelling in heavily congested conditions, riding on an overcrowded bus, or travelling with a friend, leads to major differences in perception and reporting of travel time. Negative content of time leads to overestimates of time and distance, while positive content leads to underestimates. In addition, depending on the nature of the survey, people may intentionally misreport times or distances to justify some perceived value judgement in the survey. Most transport planners would like to collect route-choice data. However, those that have attempted this task have almost invariably failed. A few surveys have been able to determine freeway use on a trip. The lack of empirical data on route choice is one reason that most commonly used assignment procedures are based on the incorrect paradigm of minimising travel time. Recently, GIS applications in transport have enabled attempts to deal with the problem of unknown trip-end addresses by equipping interviewers with computerised gazetteers. These allow them to probe subjects intelligently about the locations of trip starts and ends. However, the up-front costs of compiling such a gazetteer are large compared to the data-collection task. This method has had only limited success in dealing with the geographic location issue. It does not deal with the issues of accurate times, distances, or routes.

This paper presents the application of fuzzy-neuro-evolutionary hybrid system with online reinforcement learning for intelligent road traffic management and control. The hybrid system presents different methodologies in knowledge acquisition, decision-making, learning and goal formulation with the use of a three-layered hierarchical, distributed agent architecture. Distributed and hierarchical fuzzy knowledge acquisition allows different levels of perception to be derived for the same traffic situation by the intelligent agents. Agents' perceptions can be changed with the use of online reinforcement learning. Initial experimental results show that the implementation of the hybrid agents in a signalized network generally yields better traffic performance when compared to a network without the agents. The probability of a traffic network evolving into pathological states with oversaturation is also reduced with the implementation of the agents.