Transportation Research Record: Journal of the Transportation Research Board

Surface terrain information is needed to economically site new or relocate existing infrastructure facilities and make final design plans. Field surveying and photogrammetric mapping are the methods most widely used to acquire these data. However, these methods are time- and resource-intensive, as significant data collection and reduction are needed to provide the level of detail necessary for facility location and design. Light detection and ranging (LIDAR) is a relatively new alternative technology for obtaining terrain information more efficiently. With LIDAR, data can be collected under a variety of environmental conditions, including low sun angle, cloudy skies, and even darkness, resulting in expanded windows for data collection. Although less accurate than photogrammetric mapping, LIDAR can help expedite the highway location and design process by providing designers with preliminary terrain information earlier in the process. Presented is a proposed methodology for using LIDAR in conjunction with photogrammetric mapping to speed up highway location and design activities, including estimates of time and cost savings.

An application of light detection and ranging (LIDAR) technology to highway intersection safety is presented. LIDAR can be used to collect information about a surface by reflecting thousands of light beams per second off the surface and measuring the return time of the beams. The surface profile is collected as a digital signature that can be used in a variety of applications. Collection of information on the surface profile of the earth in the form of elevation data is one of several LIDAR applications that have been used for mapping and contouring. The focus of the described application is use of LIDAR elevation data to obtain information on intersection geometry that can lead to the discovery of potential obstructions in driver sight lines. After appropriate transformations, LIDAR elevation data were used in line-of-sight analysis to obtain information on sight-line obstructions at six intersections on the IA-1 corridor in Iowa. Intersection crash frequency and data availability were considerations in the selection of the six intersections. Results from the line-of-sight analysis were validated by visits to the intersections in the field and verification of the existence of obstructions detected during the analysis. Sixty-six lines of sight were blocked during the line-of-sight analysis, of which 62 (89.8%) were confirmed during the validation process. Four (5.8%) sight-line obstructions were not confirmed during the validation. At least three (4.4%) potential sight-line obstructions discovered during validation were not detected during the line-of-sight analysis. The intersection with the highest crash frequency was correctly found to have obstructions located within the intersection sight triangles. It can be concluded that LIDAR elevation data can be used successfully for identifying potential sight-distance problems at intersections. Identified potential problems can be verified and rectified in the field. LIDAR is a relatively costly data source, and a single application, such as this one, cannot justify the high cost of LIDAR data acquisition. Other potential highway safety enhancing applications of LIDAR must be investigated to offset the high data-acquisition cost. Suggestions for other highway safety applications are provided.

The framework of a traffic prediction model that could eliminate noise caused by random travel conditions is investigated. This model also can quantitatively calculate the influence of special factors. The framework combined several artificial intelligence technologies, such as wavelet transform, neural network, and fuzzy logic. The wavelet denoising method is emphasized and analyzed.

With the fast-growing telematics market and maturing traffic-information services, telematics devices provide a feasible means with which to manage traffic more efficiently. The provision of traffic information to travelers usually involves different parties that have distinctive objectives: travelers are concerned with benefits of travel-time savings at an affordable service charge, private information service providers (ISPs) seek to provide marketable information services from which they can derive a profit, and traffic management centers (TMCs) have the responsibility to maintain and improve system performance, especially to minimize the total system travel time. How transportation system managers can analyze the trade-offs among these objectives and adjust this new traffic-information flow diagram to improve system performance remains an open question. The trade-offs needed among the conflicting multiple objectives of different parties are studied, and traffic system performance is analyzed. The complex traffic network is formulated as a bilevel program. The upper level can be formulated by using various objective functions, such as the objectives for ISP and TMC. The lower level is a multiclass dynamic traffic-assignment model, which determines dynamic traffic flows in the network by considering the information dissemination strategies provided by the upper-level model. Numerical results of a small network are provided to illustrate the behavior of this model, and they prove that when there is congestion in the dynamic transportation network, appropriate subscribed rates benefit both all travelers and system performance, while the ISPs’ information influences little without congestion in the transportation network.

A Bayesian belief network (BBN) is a modeling and knowledge-representation structure used in artificial intelligence that consists of a graphical model depicting probabilistic relationships among variables of interest. This graphical model is a valuable tool for representing the causal relationships in a given set of variables. Because the number of possible BBNs for a given data set is exponential with respect to the number of variables, learning a BBN from data is a difficult and resource-consuming task. A greedy algorithm that automatically constructs a BBN from a data set of cases obtained from a household survey was implemented. The resulting BBN shows the dependencies among key variables that are associated with the trip-generation process.

Because major capacity-expansion projects are very unlikely in the coming years, transportation planners need to view the existing infrastructure as fixed and to 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. Modular artificial neural networks (ANNs) were developed for solving the inverse transportation planning problem. ANNs were designed to predict zonal trip ends, given the traffic volumes on the links of the transportation network. Computational experiments were performed to study the effect on ANN accuracy of three factors: transportation network size, variability in training data, and ANN topology. ANNs were shown to be quite capable of capturing the relationship between link volumes and zonal trip ends for both small and medium-sized transportation networks and for degrees of variability in the training data. Modular ANNs with one or two hidden layers appeared to outperform other ANN topologies.

An approach is presented for optimizing short-term traffic-prediction performance by using multiple topologies of dynamic neural networks and various network-related and traffic-related settings. The conducted study emphasized the potential benefit of optimizing the prediction performance by deploying multimodel approaches under parameters and traffic-condition settings. Emphasis was placed on the application of temporal-processing topologies in short-term speed predictions in the range of 5-min to 20-min horizons. Three network topologies were used: Jordan–Elman networks, partially recurrent networks, and time-lagged feedforward networks. The input patterns were constructed from data collected at the target location and at upstream and downstream locations. However, various combinations were also considered. To encourage the networks to associate with historical information on recurrent conditions, a time factor was attached to the input patterns to introduce time-recognition capabilities, in addition to information encoded in the recent past data. The optimal prediction settings (type of topology and input settings) were determined so that performance was maximized under different traffic conditions at the target and adjacent locations. The optimized performance of the dynamic neural networks was compared to that of a statistical nonlinear time series approach, which was outperformed in most cases. The study showed that no single topology consistently outperformed the others for all prediction horizons considered. However, the results showed that the significance of introducing the time factor was more pronounced under longer prediction horizons. A comparative evaluation of performance of optimal and nonoptimal settings showed substantial improvement in most cases. The applied procedure can also be used to identify the prediction reliability of information-dissemination systems.

While information technology has facilitated the collection of neverbefore-seen quantities of data, these data have not always provided the information needed by transportation professionals to support sound decision making. Computational intelligence (CI) has great potential to support the needs of transportation professionals. CI is a result of synergy among information processing technologies such as artificial neural networks (ANNs), fuzzy sets, and genetic algorithms. As the number of CI applications to transportation problems grows, so does the need to evaluate these systems. The issue of validating and evaluating transportation CI applications is addressed. A case study that evaluates the effectiveness of two CI paradigms, case-based reasoning and ANNs, for estimating the benefits of real-time traffic diversion is presented. The case study illustrates the need for regarding validation and evaluation as a part of the development effort and the need for tuning the design parameters of CI paradigms.

A multiagent architecture for real-time coordinated signal control in an urban traffic network is introduced. The multiagent architecture consists of three hierarchical layers of controller agents: intersection, zone, and regional controllers. Each controller agent is implemented by applying artificial intelligence concepts, namely, fuzzy logic, neural network, and evolutionary algorithm. From the fuzzy rule base, each individual controller agent recommends an appropriate signal policy at the end of each signal phase. These policies are later processed in a policy repository before being selected and implemented into the traffic network. To handle the changing dynamics of the complex traffic processes within the network, an online reinforcement learning module is used to update the knowledge base and inference rules of the agents. This concept of a multiagent system with online reinforcement learning was implemented in a network consisting of 25 signalized intersections in a microscopic traffic simulator. Initial test results showed that the multiagent system improved average delay and total vehicle stoppage time, compared with the effects of fixed-time traffic signal control.

Trip-production rates presented in cross-classification tables are essential data for the planner’s understanding of the travel characteristics of a region. Trip rates obtained from surveys, however, often show a pattern that is not consistent with what is expected by the analyst; for example, the greater the household size and auto ownership, the greater the number of trips generated. This pattern may not be found in the trip rates that are obtained directly by the survey. In such cases, analysts commonly adjust the irregularities manually. The way in which the values are adjusted affects the credibility of the trip table and, ultimately, the forecast travel demand. A method that adjusts the values of the trip table systematically is presented. The process uses the fuzzy linear programming method. The objective is to make the adjusted value as close to the observed value as possible. The constraints are to make the adjusted values adhere to the analyst’s general expectations about the pattern of the values in the table, and to match the number of trips estimated from the adjusted trip table with the actual number of trips surveyed. An application example that uses real-world data is given.

The CUSUM (cumulative sum of log-likelihood ratio) algorithm is an optimization-based algorithm that is attractive for many applications because it can minimize detection delay and can explicitly incorporate the characteristics of processes before and after changes. One such application is freeway incident detection, where field-measurable traffic-flow parameters are used to flag incidents in real time in an expedient and reliable manner. In the presented study, the special characteristics of traffic processes associated with incidents are incorporated into the CUSUM algorithm for freeway incident detection. In the algorithm evaluation, the most recently developed neural networks are compared with an enhanced CUSUM algorithm. The neural network algorithms are systematically evaluated first among themselves, and then the best of them is compared with the CUSUM algorithm. The results demonstrate that the CUSUM incident detection algorithm can perform better than the neural network algorithms. The neural network algorithm may show inferior performance because it cannot adjust its decision threshold in real time.

Creating and maintaining up-to-date, sharable, geographic information systems–transportation data are challenging. Many states are working on transportation framework efforts to build a complete, consistent, and current transportation data layer in conjunction with the National Spatial Data Infrastructure and Geospatial One-Stop efforts. An effort was made to systematically reexamine transportation data-sharing issues as part of the development of a transportation framework for the state of Washington. Business needs were assessed for spatial and temporal accuracy needs of stakeholders and users of a statewide transportation framework. A conceptual model was developed for a transportation framework with emphasis on data flows to update and maintain it and other transportation databases. An update clearinghouse is proposed to collect and disseminate data about new and modified road features. A synthesis of issues is presented, including determining the desired robustness of the transportation framework and cost-allocation principles for possible add-on enhancements.

The Global Positioning System (GPS) offers a cost-effective and accurate method for inputting, updating, and analyzing spatial data. However, many data-storage and reporting methods at state departments of transportation and other transportation agencies rely on linear referencing methods (LRMs) for managing transportation data. Consequently, GPS data must be able to coexist with linear referencing systems (LRSs). Unfortunately, the two systems are fundamentally different in the way they collect, integrate, and manipulate data. For spatial data collected with GPS to be integrated into an LRS or shared among LRMs, several issues must be addressed. Various issues are discussed for integrating point features from a typical inventory system with an LRM or between LRMs, including two- or three-dimensional GPS to one-dimensional LRM, linear offset error, lateral offset error, matching to the wrong segment, and locating points between LRMs.

Intelligent transportation system (ITS) sensor networks, such as road weather information and traffic sensor networks, typically generate enormous amounts of data. As a result, archiving, retrieval, and exchange of ITS sensor data for planning and performance analysis are becoming increasingly difficult. An efficient ITS archiving system that is compact and exchangeable and allows efficient and fast retrieval of large amounts of data is essential. A proposal is made for a system that can meet the present and future archiving needs of large-scale ITS data. This system is referred to as common data format (CDF) and was developed by the National Space Science Data Center for archiving, exchange, and management of large-scale scientific array data. CDF is an open system that is free and portable and includes self-describing data abstraction. Archiving traffic data by using CDF is demonstrated, and its archival and retrieval performance is presented for the Minnesota Department of Transportation–s 30-s traffic data collected from about 4,000 loop detectors around Twin Cities freeways. For comparison of the archiving performance, the same data were archived by using a commercially available relational database, which was evaluated for its archival and retrieval performance. This result is presented, along with reasons that CDF is a good fit for large-scale ITS data archiving, retrieval, and exchange of data.

Dynamic segmentation is viewed as one of the most important functions of geographic information systems for transportation applications. Although the road network and associated events (e.g., pavement material, traffic volume, incidents) can be referenced to both space and time, the spatial and temporal dimensions have not been well integrated. Modeling space-varying, time-varying, and space-time-varying events in dynamic segmentation by using an object database approach that is in line with the Object Database Management Group standard is explored. A mechanism called parametric polymorphism is used to lift conventional data types to spatial, temporal, and spatiotemporal types for maintaining knowledge about events that could change spatially, temporally, and spatiotemporally along linear features. An associated object query language, DS-OQL, was designed to support the formulation of spatial, temporal, and spatiotemporal queries on the road and event information.

Given the enormous quantities of data collected by intelligent transportation systems (ITS), transportation professionals recently have focused on developing archived data user services (ADUS) to facilitate efficient use of these data in myriad transportation analyses. Most ITS systems were designed by using a transactional database design, which is not well suited to support the ad hoc queries required by ADUS. Research investigated the application of data-warehousing concepts to better support the requirements of ADUS. A case study is presented in which an ADUS for the Hampton Roads Smart Traffic Center, the regional freeway management system, was redesigned from a transactional approach to one based on data warehousing. Test queries run by using both approaches demonstrated that dramatic increases in efficiency are achievable through the use of data-warehousing concepts in ADUS.

Many states have implemented large-scale transportation management systems to improve mobility in urban areas. These systems are highly prone to missing and erroneous data, which results in drastically reduced data sets for analysis and real-time operations. Imputation is the practice of filling in missing data with estimated values. Currently, the transportation industry generally does not use imputation as a means for handling missing data. Other disciplines have recognized the importance of addressing missing data and, as a result, methods and software for imputing missing data are becoming widely available. The feasibility and applicability of imputing missing traffic data are addressed, and a preliminary analysis of several heuristic and statistical imputation techniques is performed. Preliminary results produced excellent performance in the case study and indicate that the statistical techniques are more accurate while maintaining the natural characteristics of the data.

The traffic-forecasting model, when considered as a system with inputs of historical and current data and outputs of future data, behaves in a nonlinear fashion and varies with time of day. Traffic data are found to change abruptly during the transition times of entering and leaving peak periods. Accurate and real-time models are needed to approximate the nonlinear time-variant functions between system inputs and outputs from a continuous stream of training data. A proposed local linear regression model was applied to short-term traffic prediction. The performance of the model was compared with previous results of nonparametric approaches that are based on local constant regression, such as the k-nearest neighbor and kernel methods, by using 32-day traffic-speed data collected on US-290, in Houston, Texas, at 5-min intervals. It was found that the local linear methods consistently showed better performance than the k-nearest neighbor and kernel smoothing methods.

The Dempster–Shafer data-fusion technique as affected by probability masses as a result of sensor selection and probability masses distribution is investigated. Dempster–Shafer inference is a statistically based data-classification technique for detecting traffic events that affect normal traffic operations. It is used when data sources contribute discontinuous and incomplete information such that no single data source can produce a predominantly high probability of certainty for identifying the most probable event. To help in the selection of appropriate sensors and probability masses, a rough-sets data-mining technique in support of Dempster–Shafer inference was proposed and tested. The basic rough-sets technique is introduced, and a numerical example is given to explain its applications. Field testing of the rough-sets technique showed that it can reasonably and systematically process a large amount of traffic information, as an alternative to relying on the intuition of traffic operators and system managers. Because it allows easy maintenance and update of estimated probability masses, this technique is suitable for large-scale applications at the traffic management center.