8. CONCLUDING SUMMARY

            Intelligent Transportation Systems (ITS) are an alternative data source that could conceivably lead to win-win situations. ITS-generated data will not only benefit the transportation operations and planning communities by allowing them to access more and better data. It will also enhance the appeal of ITS deployment by significantly broadening its originally intended benefits. The notion of using ITS-generated data as an alternative data resource is reflected in the Archived Data User Services (ADUS) in the National ITS Architecture.

            Usually, an agency will evaluate the costs and the benefits of ADUS before it decides whether to deploy ADUS. The cost of archiving and using ITS-generated data is typically measured in terms of the effort needed to archive and re-format the data, re-vamp the software, and address data quality and data integration issues. The benefits, on the other hand, are measured in terms of the value added by the ITS-generated data. Although at this point the costs are high to use ITS-generated data for purposes other than the originally intended use, our previous research^8.1 has proven the concept that ITS-generated data can indeed improve transportation decisions by, in this particular case, improving traffic estimates.

8.1 BENEFITS OF ADUS

            In the past, transportation planning, operations and preservation were constrained by information that was typically out-of-date and limited in scope. That situation arose because data collection is expensive, and information compilation and dissemination take time. Data limitations have been widely and duly recognized, and addressing this limitation is one of the themes in the forthcoming reauthorization legislation.

            With the recent advent of intelligent transportation systems, these systems provide another source of data. Archived ITS-generated data are distinct from traditional data sources in three aspects: (1) ITS-generated data are temporally intensive (e.g., collected in very short intervals), (2) ITS-generated data meet some major data gaps that could not be met in the past due to resource limitations, and (3) all ITS-generated data are on electronic media, thereby expediting data analysis and information dissemination.

            These attributes of archived ITS-generated data provide unprecedented opportunities that traditional ways of compiling information can not offer. Making information accessible almost on a real-time basis allows transportation planners and operators to anticipate emerging issues, thereby allowing them to progress from a reactive mode to a proactive mode. Furthermore, more detailed and insightful understanding of the problems (safety, planning, operations, or maintenance) is now possible because of the expanded scope and increased frequency of data collection.

            The specific benefits of using ITS-generated data vary from one application to the next, and are difficult to enumerate. However, the general benefits of using archived ITS-generated data can be gauged in at least three ways. First, can ITS data replace traditional data? If so, this benefit can be measured in monetary terms. Second, can ITS-generated data supplement traditional data so that more reliable estimates can be developed? Third, can ITS-generated data meet data gaps that are expensive or impossible to meet with traditional data sources? Results from our previous research confirm that ITS-generated data can both replace and supplement data collected through traditional ways^8.2.

            As more and more ITS is deployed in the future, ITS-generated data can replace many of the current data collection processes. However, before then, the greatest contribution that ITS data can offer is probably when they are integrated with, or used to supplement, traditional non-ITS data to address gaps in these data.

8.2 BARRIERS TO ADUS

            ADUS can be viewed as a progression in three phases with increasingly challenging activities: archiving data, using the archived data, and sharing the archived data. Consequently, the barriers to ADUS are characterized into three categories: (1) barriers to archiving data, (2) barriers to using archived data, and (3) barriers to sharing archived data. Although similarities exist, the barriers and benefits of each of these phases can be considerably different from those in other phases,. The decision to archive and use ITS-generated data typically hinges on the trade-off between the costs of removing the barriers and the benefits from using the archived data.

            The nature of these three types of barriers can be characterized as decreasingly technical and increasingly institutional. For example, the barriers to archiving data are typically more technical in nature, while the barriers to sharing archived data are largely institutional. Ideally, when barriers in all three phases are removed, ADUS would reach its greatest potential. Nonetheless, the benefits are still substantial even when barriers in only a single phase are overcome.

            Overall, ADUS barriers include institutional inertia, concerns over privacy, proprietary concerns, liability, data ownership, data liability, data integrity and quality, data compatibility, and other technical and technological issues. They can be categorized into five areas:

            ●    Institutional impediments,

            ●    Data issues,

            ●    Lack of standardization,

            ●    Privacy and liability, and

            ●    Other technological barriers.

            In addition to the direct costs, the costs include efforts needed to:

            ●    Articulate and communicate the needs for archived data,

            ●    Save the data,

            ●    Re-format the archived data into a user-friendly format,

            ●    Re-vamp the existing software to accommodate archived data,

            ●    Address data quality issues,

            ●    Make archived data accessible in a timely manner,

            ●    Integrate the archived data with non-ITS data to meet broader data needs,

            ●    Reconcile data incompatibilities among different data sources, and

            ●    Forge mutually beneficial partnerships among data-producing agencies and data users.

8.2.1 Barriers to Archiving Data

            Results from the ITS Deployment Surveys, case studies and the literature review suggest that the level of data archiving and use varies from one application to the next, and from one locale to another. Transit applications are significantly more advanced and widespread than the other three applications. The prevalence of the transit community in using ITS-generated data seems to stem from the nature of its organizational structure. The data “producers” are oftentimes the data “users.” This structure greatly reduces the institutional barriers that challenge the other three user communities.

            The planning community is not progressing as rapidly as the transit community on ADUS. This is primarily due to two reasons. First, the entities that produce the desirable planning data (e.g, the TMCs) are organizationally separate from the data users (e.g., the planners). Second, many data producers do not currently see the need to archive their data. When more data producers experience the benefits of the archived operational data, more planning applications will likely be implemented.

            Highway safety is perhaps the area that has been targeted the least in terms of data archiving. There are two contributing factors. First, the implementation of ADUS, as it relates to highway safety, faces an unique problem. Of all the data elements collected by ITS deployments, none of them collect data that are strictly about safety. Traffic management centers collect data in order to keep traffic flowing smoothly, which may benefit safety, but safety remains tangential to their work. Those who are interested in safety have to find their own ways to relate ITS-generated data to the topic of highway safety.

            Another safety issue is the heightened concern over protecting the identity of incident victims. This heightened sensitivity can slow down or limit the flow of certain types of data. Apart from the barriers common to data archiving in general, archiving data pertinent to highway safety is confronted by a very different challenge. Part of the traffic information on freeways and arterial roads is collected by cameras. In particular, video cameras are used to identify the exact locations and circumstances where something affects traffic on the freeway system. That type of information is enormously valuable in advancing our understanding of crash causation so that effective countermeasures can be deployed. Notwithstanding these potential benefits, access to that information has been extremely controversial. Undoubtedly, the biggest barriers to archiving data that could have some bearing on highway safety applications are privacy and liability issues. Recognizing these concerns, many agencies limit their data archiving or withhold safety information and camera images. On limited occasions, images are recorded for traffic studies (such as vehicle counts, weaving movements), training purposes, and exceptional circumstances (such as in criminal investigations)^8.3.

            The second contributing factor, that limits the application of ITS-generated data to improve safety, is the complexity of highway safety applications. Safety is a complex interaction among vehicles, roadways, environments, and drivers. By the same token, highway safety could be the area that will benefit the most from data archiving. Since highway safety transcends geographic boundaries (metropolitan vs. rural), data archived from almost any of the ITS infrastructure components can be used to meet some of the data requirements in highway safety applications. For example, volume data collected from an ITS traffic surveillance project can be used to estimate accident exposure by time of day and day of the week. Another example is archived speed data which, in conjunction with information on highway geometry and weather, can be used to estimate the propensity of incidents or accidents. That said, data integration becomes ever more imperative in highway safety applications.

            Across all of the applications, another major barrier to archiving data is the question of who should bear the cost of data archiving--data providers such as Transportation Management Centers (TMCs) or data users? In an increasing number of cases, this cost has been borne by the data user. For example, data are made available by data brokers on the internet and are accessible by the user through subscriptions. Or, agencies who recognize the cost effectiveness of using the archived data (i.e., data user) provide funding to the data producers (e.g., TMCs) so that ITS-generated data can be archived, and sometimes reformatted, to meet the data user’s specific needs. Alternatively, the data “users” contract to a third party to archive and analyze the ITS-generated data.

8.2.2 Barrier to Sharing Archived Data

            Our findings also suggest that different applications face different concerns in sharing archived data. Data sharing becomes almost impossible when it risks violating individuals’ privacy rights. Although the privacy concern varies from one agency to the next, there is a clear need to investigate the use of privacy protection technologies, such as cryptographic technologies, to remove these concerns.

            Other than the resistance within an agency to change or to adapt to new data systems, the traditional stove-pipe organizational structure also discourages data sharing among agencies. Furthermore, this disinclination to share data among agencies involves a number of non-institutional issues. For example, the possibility of sharing data with sister agencies raises concerns about one’s data quality –“Are we comfortable enough to share our data with others?” and “Who is responsible for data quality assurance?” This concern about data quality is an unexpected “barrier” to information sharing. On the other hand, it might be an unintended benefit of ADUS deployments because agencies might be spurred to improve the quality of their data before they share them. Before ADUS can be widely adopted, this data stewardship/ownership issue needs to be addressed. Incompatibility of data formats, data models, and analysis software also hinders data sharing among agencies.

            The lack of communication or the lack of awareness adds another barrier to sharing and using ITS-generated data. In many cases, potential users of archived data are unaware of the existence of these data collections. For example, officials in City X are aware of the data collected in their own area, but they may not be aware of the data collected in City Y. A transportation engineer in City Z might be unaware of either collection of data. This lack of awareness about the availability of archived data is a very real barrier.

            Proprietary rights to value-added data (as such in the cases of freight and CVO applications) can be another barrier to sharing the data. A lack of institutional cooperation among various data users might mean that disparate users duplicate data collection rather than jointly share in the costs of collections, ownership, and use.

8.2.3 Barriers to Using Archived Data

            Data collection is an extremely expensive undertaking. As such, the purposes of any data collection efforts are usually clearly defined and understood before the data collection begins. Even more desirable would be a comprehensive plan, before data collection, of how the data should be analyzed after they have been collected so that the purposes of the data collection effort will be fully addressed. This is the typical principle of any data collection effort. Examples of such data collections include household travel surveys, commodity flow survey, highway monitoring programs, the Highway Performance Monitoring System (HPMS), and the General Estimates System (GES).

            Limited resources sometimes lead to the alternative of leveraging on data recorded/reported for administrative purposes. Examples include the International Registration Plan (IRP), International Fuel Tax Agreement (IFTA), drivers’ licenses, vehicle registrations, police accident reports, etc. Attempting to use data to meet data needs for which data are not originally intended is significantly more challenging than using data for purposes for which the data are originally intended. Since the notion of ADUS is to use ITS-generated data in applications other than what is originally intended, ADUS data are in this category.

            The difficulty to use archived data is confirmed by a common reaction from data producers and data users: “...the benefits of using archived data to meet my data gaps are obvious, but it is not at all clear how I can use these data and for what.” This barrier could be overcome by generating and disseminating lessons learned, developing “how-to” guidebooks, deploying field tests, and developing easy-to-use tools on calculating the expected range of costs and benefits of different ADUS applications.

8.3 OPPORTUNITIES

            Findings from the case studies in this project suggest that the advantages of archiving and sharing archived ITS-generated data are recognized and capitalized on by almost all stakeholders. That said, some stakeholders remain somewhat skeptical in that the benefits of ADUS have to be clearly demonstrated in terms of costs and benefits before any consideration will be given to ADUS applications. The most common hurdles are: “It [ADUS] sounds really nice, but there is not adequate staff and time to figure out what problems can be addressed by the archived data.” or “What is in it for me?” To overcome these barriers, the benefits of ADUS have to be clearly articulated through field testing that must be carefully crafted to address very specific issues.

            Rather than try to identify all possible opportunities, this project identifies those that are practically feasible, can be quickly deployed, and are most likely to produce immediate benefits/results. The rationale for identifying these “low-hanging fruits” is that the sooner that quantifiable benefits of using ITS-generated data for operations or planning improvement are demonstrated and disseminated, the sooner additional deployments will be stimulated. These opportunities are summarized in Table 8.1. Background information and objectives on each of these opportunities are discussed in the respective chapters. Note that a few of them are cross-cutting in that they are pertinent to multiple applications. For example, work zone safety can be a collaborative effort between the operations and the safety communities.

            Opportunities that are practically feasible, can be quickly deployed, and are most likely to produce immediate benefits/results are those that are able to:

          ●    identify specific technical and institutional barriers to archiving, using, and sharing ITS-generated data;

●    develop feasible solutions to overcome these barriers;

           ●    identify issues pertinent to standards development;

           ●    examine the feasibility of integrating ITS-generated data with data collected from traditional and emerging technologies (e.g., highway monitoring data, remotely sensed data);

           ●    identify and quantify costs and benefits;

           ●    disseminate lessons learned, and

           ●    share the developed procedures and software in an open-source environment. Some examples of these procedures and software are: those developed to convert raw ITS-generated data into formats acceptable to existing and/or off-the-shelf data management or analysis software, check the quality of the data, impute missing data, correct questionable data, abstract information suitable for data analysis from “text” files, estimate potential recurring and non-recurring traffic delays, and other applications. The benefit of sharing these procedures and software in an open-source environment is that it reduces the “re-inventing the wheel” thus enabling more efficient use of resources.

            Although the costs are high at this point to use ITS-generated data for purposes other than the originally intended use, many ADUS applications have demonstrated that ITS-generated data can indeed improve transportation decisions by improving the quality of the existing data or by meeting unmet data needs. As more and more ITS is deployed in the future, ITS-generated data can no doubt replace many of the current data collection processes. However, before then, the greatest contribution that ITS data can offer is probably when they are integrated with, or used to supplement, traditional non-ITS data to address gaps in these data.

            Despite the promise of ITS-generated data, it is imperative to caution that ITS-generated data should not be viewed as a “silver bullet” for addressing data gaps. Furthermore, it can be extremely misleading to assume that ITS data are “ready-to-use.” Extensive and thorough data quality checks which can be extremely demanding are essential before ITS-generated data can be used.

            It is also important to point out that the costs of using ITS-generated data should decrease considerably as more uses are stimulated. At that point, the costs could very well become inconsequential compared to the benefits. Additional research and analysis would be useful to quantify these costs and benefits, to help standardize the data archiving, to develop standards, to improve methods to “prepare” and use ITS-generated data, and to develop simplified user-friendly analysis tools.

ENDNOTES:

Return to Table of Contents