4
Data Sources, Models Used, and Emerging Technologies

POTENTIAL DATA SOURCES

Because there are no direct measures of ton-miles and value-miles of international trade traffic carried by highways for each state, data from a variety of sources, none of which was specifically developed for this purpose, have to be used to provide estimates. Previous National Research Council studies have assessed several of these data sources (see, e.g., National Research Council, 2003a).

The congressional legislation that mandated the Bureau of Transportation Statistics (BTS) study implicitly recognized many of the difficulties that would be encountered in developing measures of ton-miles and value-miles of international trade traffic carried on highways in each state, and it directed BTS to identify long-term data improvements to provide accurate and reliable measures for use in highway apportionments. This section provides a few observations about the potential of various approaches to sufficiently improve data to enable its use in formula allocation.

A prior BTS report (Bureau of Transportation Statistics, 2003) listed four general deficiencies in the current data sources: (1) there are international shipments involving Canada or Mexico that are not exports or imports;1 (2) short-haul shipments are not included; (3) information on all

1  

We note that this deficiency may not be due to data limitations—it may be definitional.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 20
Measuring International Trade on U.S. Highways 4 Data Sources, Models Used, and Emerging Technologies POTENTIAL DATA SOURCES Because there are no direct measures of ton-miles and value-miles of international trade traffic carried by highways for each state, data from a variety of sources, none of which was specifically developed for this purpose, have to be used to provide estimates. Previous National Research Council studies have assessed several of these data sources (see, e.g., National Research Council, 2003a). The congressional legislation that mandated the Bureau of Transportation Statistics (BTS) study implicitly recognized many of the difficulties that would be encountered in developing measures of ton-miles and value-miles of international trade traffic carried on highways in each state, and it directed BTS to identify long-term data improvements to provide accurate and reliable measures for use in highway apportionments. This section provides a few observations about the potential of various approaches to sufficiently improve data to enable its use in formula allocation. A prior BTS report (Bureau of Transportation Statistics, 2003) listed four general deficiencies in the current data sources: (1) there are international shipments involving Canada or Mexico that are not exports or imports;1 (2) short-haul shipments are not included; (3) information on all 1   We note that this deficiency may not be due to data limitations—it may be definitional.

OCR for page 20
Measuring International Trade on U.S. Highways modes of shipment using multiple modes is often missing; and (4) complications can result from situations in which the country of origin can change when the goods are substantially transformed prior to end use. For data on exports, the BTS report also identified five deficiencies in the Commodity Flow Survey (CFS): (1) it excludes a number of industries; (2) it has substantial nonresponse and sampling variance for exports; (3) it is only collected every 5 years; (4) for over 30 percent of its export records, the port of exit is missing; and (5) it does not provide useful information on imports. For data on imports, the BTS study listed five deficiencies: (1) the value of imports must be estimated from commodity code and weight; (2) truck traffic from maritime imports must be estimated by the total minus imports transported by rail, which ignores petroleum products and inland and coastal shipments by water; (3) allocation to the state level uses Port Import/Export Reporting Service (PIERS) data, for which the state of destination is often missing; (4) the state may be the address of the importer and not that of the destination; and (5) truck traffic from imports arriving by air is ignored. Finally, the BTS report pointed out two problems in estimating transportation at the substate level: there is datedness and a lack of focus on foreign business in the County Business Patterns data, and there are missing values of various kinds in calculating the Oak Ridge National Laboratory’s (ORNL) highway network and the intermodal network. To address these deficiencies, the BTS report listed five data needs: (1) data on shipment weight for imports transported by truck, to eliminate the need to estimate this by subtraction; (2) information on state and county of destination for imports and information on county of destination for exports, to eliminate the need to use PIERS data and the county allocation model; (3) information on the port of entry rather than the administrative port; (4) data on all modes of transportation used in shipping to a destination, in addition to the mode used when the cargo arrives in or departs from a U.S. port of entry; and (5) data on in-transit and transshipments. The BTS report mentioned three possibilities for (additional) data collection that we review in this section: (1) obtaining access to currently restricted data; (2) expanding the information requested on administrative records; and (3) initiating a new survey. In addition, there are a number of emerging technologies that may be useful for this purpose; we consider them in the last part of this section.

OCR for page 20
Measuring International Trade on U.S. Highways Obtaining Access to Currently Restricted Data The Census Bureau interprets Title 13 provisions on data privacy to extend to data from the Customs and Border Protection Service. We are not fully aware of the underlying reasons for this ruling but it may be that some compromise can be arrived at through one of the following avenues: (1) requesting use of an extract at some aggregate level or on a sample basis with all identification of the ownership omitted to help ensure privacy; (2) using various analytic techniques for disclosure protection, such as noise incorporation; (3) providing the software for analysis to the Census Bureau and requesting that the Bureau provide the aggregate estimates at the county level in return, observing disclosure protection; or (4) having BTS personnel work in the Census Bureau as special sworn employees in a restricted-access environment. Expanding the Information Requested on Administrative Records The possibility that PIERS data could include collection of data on shipment value and the state of destination rather than the state of the importer should be explored. In addition, the cause of missing data on state of destination should be examined and actions taken to reduce the occurrence of missing data. International trade statistics are collected by the Customs and Border Protection Agency for most U.S. international transactions and are processed by the Census Bureau’s Foreign Trade Division. This database is large, containing more than 3 million import and 2 million export transactions per month (Monk, 2003). If the records in this database included a few additional key data elements, they could be used directly to report international trade volumes, values, destinations, and modes. However, the data are largely administrative in nature, provided by shippers, not carriers. The records contain value data but no weight data; they have limited information on geographic location of origin and destination; they have uncertain assignment of port geography; and they fail to capture all modes of transportation. The export statistics are of particularly questionable quality since they receive little agency scrutiny. Estimates suggest that undercoverage of exports is between 3 and 10 percent. All of these factors tend to limit the utility of these data for apportionment formula purposes. The panel observes that there are several fortuitous projects under way, which, in the long term, hold promise of filling these information gaps.

OCR for page 20
Measuring International Trade on U.S. Highways The Custom and Border Protection Agency is streamlining and enhancing the international transaction data systems under the Automated Commercial Environment (ACE) system, a $1 billion program. This upgrade is linked to a government-wide program that would meet the needs of more than 100 agencies for trade data. Over the next several years, more than $100 million will be invested in the International Trade Data System (ITDS) to modernize the data and access to them (Fiocco, 2003). This combined effort holds promise, in the long term, of filling information gaps relevant to the estimation of international highway trade traffic. The ITDS could provide a “one-stop” filing system for both carriers and shippers through use of an Internet-based integrated government-wide system of collection of trade and transportation data. Such a system could meet the needs of all federal agencies with responsibilities related to international trade. Under ACE and ITDS data will be reported from both carriers and shippers, with a unique identifier associated with each international transaction. The plan is for ITDS/ACE to become the government’s front-end information technology system for all federal trade and border agencies, providing more accurate, thorough, and timely data on imports, exports, and in-transit information on cargoes, conveyances, and crews. The hope is to use these data to improve compliance with and enforcement of trade requirements, reduce the cost and burden of processing trade transactions, facilitate interagency coordination, develop and distribute higher quality and more comprehensive and timely information on U.S. trade, and improve Department of Transportation access to this information. Integration is extremely desirable since there are more than 100 agencies and offices involved in trade, and as a result, some data collection has been redundant, and data collection currently makes use of some paper forms and incompatible automated systems. ITDS and ACE will also promote standardization of federal international trade and transportation requirements. There is certainly a tension between information and security, so it is not surprising that while ITDS and ACE may have the promise of resolving most of the data gaps discussed in this report, access may be restricted. The panel hopes that the proposed strategies for dealing with Title 13 issues concerning current Customs and Border Protection Agency data can be addressed in time to make use of this exciting new data collection and coordination effort.

OCR for page 20
Measuring International Trade on U.S. Highways Expanding the CFS or Initiating a New Survey New survey data collections are expensive, and must therefore be very carefully considered, especially given that there is a relevant survey, the Commodity Flow Survey (CFS), that is already in place. Since the CFS samples domestic exporters, one might consider augmenting the CFS with a sample of businesses that import into the United States. However, there would be substantial complications in developing a relevant frame, and the mandatory status of the CFS cannot be extended to foreign importers. A pilot study allowing analysis of the quality of the information collected might be informative and relatively inexpensive. In addition to adding information on imports, the CFS might also be examined to see whether the addition of more contextual information to address some of the data deficiencies discussed above could be accommodated without affecting the quality of the data already collected. Barring changes to the CFS, instituting a new survey raises the possibility of including data elements that are not required by the Customs and Border Protection Agency (and therefore will not be available in the future on ITDS/ACE). Two options for such a survey are a shipper-based survey or a carrier-based survey. These surveys would need to provide timely data, which may mean an annual survey, though that remains to be determined through analysis of pilot data. As noted by BTS (U.S. Department of Transportation, 2003), a carrier-based survey might involve surveying a sample of trucks entering or leaving major seaports and airports and therefore possibly engaged in international trade. (Unfortunately, carriers may not know the final destination of the cargo they carry.) The resulting data would need to be reconciled with port-specific reporting of international freight cargo. As noted in that report, in the mid-1970s the Census Bureau had access to all foreign trade paper documents collected by the Customs and Border Protection Agency, and those documents served as a universe from which a follow-up survey of shipper-based establishments was conducted. Today, an analogous survey would involve sampling international transactions filed with that agency through ACE. These surveys collected reliable data on inland destinations of imports and origin of exports, the mode of transportation used, and the weight of the cargo. Those follow-up surveys have not been conducted since the mid-1970s.

OCR for page 20
Measuring International Trade on U.S. Highways MODELS USED Three models were used in the production of the BTS estimates of ton-miles and value-miles of international highway freight: the gravity model, a state-to-county allocation model, and a road selection model. We provide short descriptions of these models and include some discussion of problematic aspects of each modeling application. Gravity Model The model initially selected by BTS to “carry down” state-level estimates of value-miles and ton-miles to the county level is known as the gravity model. A gravity model is a method for estimating the flow of goods or persons from one location to another. There are many different versions of the gravity model. The one used for the BTS analysis has the form Tij =AiBjFij, where Tij is the flow (measured, for example, in tons or value per year) from state i to state j, Ai is a state-specific parameter that represents the total tons or value originating in state i regardless of destination, Bj is a state-specific parameter that represents the tons or value arriving at state j regardless of origin, and Fij represents the “impedance” to flow between i and j. Fij is typically a composite of the cost and travel time required to ship a commodity from i to j. In the model used for the BTS study, where (k = 1,2,…, K) represents impedance-related variables, such as travel time and cost, and the θk’s are constant parameters. The Ai’s, Bj’s, and θk’s are estimated by fitting the model to data on flows between states (see Metaxatos, 2002 for further details). Because in the BTS application the model was estimated at the state level, it cannot be used directly to estimate flows at a finer level of geographical aggregation (e.g., county-to-county flows). Although the gravity model could have been applied at the level of county centroids, it would have required that the factors related to the costs of transportation from one port or border crossing or county to another be used, which was computationally too demanding (involving a matrix of dimension of more than 3,000 by 3,000). Therefore, the gravity model was applied at the level of states, which is the level at which the estimates already existed.

OCR for page 20
Measuring International Trade on U.S. Highways The panel is unconvinced that the gravity model plays a useful role in the estimation of ton-miles and value-miles for international trade traffic by highways. Replacing the observed measures with their estimated expected values (given the gravity model) in some applications could reduce variances without appreciably increasing bias. But to do so, the model form would have to be justified either by subject-matter considerations or empirically, or preferably both, since otherwise model misspecification would likely eliminate any variance-reduction benefits from the use of estimated expected values in replacing the observed measures. A second possible justification for the use of the gravity model is that one might be able to use the model for forecasting subsequent years’ highway use. However, there has been no suggestion for how the origination and destination factors could be updated for future time periods. This lack is further complicated by the introduction of new ports or border crossings over time, as well as various kinds of state or county dynamics resulting from changes in economic circumstances, user charges, various technological advances, etc. None of this is explained by the gravity model. State-to-County Share Model Using states as the geographic level of analysis does not provide very accurate assessments of distances of trade traffic, since knowing the destination state often does not specify the highways used. Therefore, it is extremely important to distribute the shipments to the level of counties (county centroid) to obtain more information about roads used. BTS used the Census Bureau’s County Business Patterns data to distribute the freight movements from the state to the county level. To do this, BTS used the ratio of each county’s total payroll to total state payroll to allocate that proportion of the state total shipments to the individual counties. (This process is adjusted to ignore establishments that do not have a county identification.) Clearly, this model is likely to be subject to substantial misspecification given its use of a single variable that is only weakly related to the transportation of freight, let alone international freight. County Centroid Route Assignment Model Given the identification of the destination county centroids, one has to identify the routes taken by trucks from seaports and border crossings to these centroids: this is accomplished through use of the ORNL highway

OCR for page 20
Measuring International Trade on U.S. Highways network assignment model, which determines the distances traveled essentially through selection of the least cost/time route. In addition, one also has to identify the mode of transportation when there are alternatives to traveling by road. One important benefit of the ORNL highway network is that it is embedded within the ORNL intermodal network, which permits one to model the mode and route used to move freight from one node in the network to another on the basis of travel costs and time required. This assignment model, within the same general approach, sometimes makes use of a least impedance approach that deterministically selects a transportation option and sometimes uses a logit approach that weights various choices based on their estimated probabilities of selection. As far as the panel is aware, neither the least impedance nor the logit approach has been supported by empirical validation. Unfortunately, this overall approach is hampered by incomplete data in several respects. In addition, there is no guarantee of the uniform applicability of the assumption that the movement of freight always conforms to an economically optimal selection of routes. The ORNL highway model produces variances, accompanying each origin-destination pair, that are estimated using the bootstrap method.2 Neither this assignment model nor its associated variances, to the panel’s knowledge, have been empirically validated with actual data on route selection. Model Validation In all three applications of models to produce the BTS estimates, there is a lack of model validation, both external and internal. External validation is the comparison of model forecasts with realized values. Internal validation is the understanding of the performance of a model through examination of the functioning of individual components, especially how variability propagates through a model. Model validation is essential both to understand the current performance of a model and to direct efforts efficiently for model improvement in the future. Model validation has benefited from enormous progress in the last 25 years, but this research is not reflected in the BTS work. (Some of the 2   The bootstrap is a technique for estimating variances in complicated situations that approximates sampling from the unknown population by sampling instead from the observed data (see, e.g., Efron, 1982).

OCR for page 20
Measuring International Trade on U.S. Highways newer methods for model validation can be found in Morgan and Henrion [1990], National Research Council [1991], and Mulry and Spencer [1991].) In an effort to validate the gravity model, BTS computed a chi-square goodness-of-fit statistic to assess how well the gravity model fits at the state level.3 However, this statistic provides little information about the performance of the model in the context of prediction. More importantly, it does not indicate whether use of the gravity model gives smaller errors in the estimated origin-to-destination flows than does use of the flow data that were used as inputs to estimation of the model. For the assignment model, ORNL has carried out a number of careful assessments of the quality of the data used to construct the networks that are relied on, but there has been no validation that the assignments produced agree with those selected in practice by freight movers, as mentioned above, and no sensitivity analyses of the effects of missing data on the assignments has been provided. In concluding this section, the panel notes one related point. Another approach to providing estimates of higher quality for states, other than additional data collection, would be to try to use combining-information models with variables that are collected more frequently than the current survey frequencies, possibly at lower levels of geographic aggregation. Making full use of any time-series structure in these independent variables might support projections several years in the future, in addition to making current small-area estimates. We have not carried any research out in this area, but we believe there is some promise for substantially improving the current estimates through use of these methods. EMERGING TECHNOLOGIES One problem with many transportation surveys is that they require time by truckers to respond, which results in substantial noncompliance 3   In the BTS study, Tij is interpreted as the expected value of the actual flow from i to j, Nij, which is treated as a random variable. Thus, Tij = E(Nij). Tij is unknown and is estimated from the fitted gravity model. Denote the estimate by Then the chi-square test statistic is with degrees of freedom equaling the total number of terms in the sum minus the number of origin factors, minus the number of destination factors, minus the number of cost factors –1.

OCR for page 20
Measuring International Trade on U.S. Highways and incomplete reporting. There are a variety of emerging technologies and techniques that might be used to more unobtrusively capture information on weights, value, and vehicle types, including weigh-in-motion and radio frequency tag technologies that would identify vehicles and cargos while in motion. There are also important advances in traffic counting technology. As described in this section, these emerging technologies that are now either in advanced development or in the early stage of deployment may soon enable more informed estimates of the amount and value of international trade traffic on state highways. The “answer” will most likely consist of a combination of technologies that provide information on both weight and cargo. It is possible to envision a system of weigh-in-motion and tandem radio frequency technologies (see below) that tie directly to emerging company and customs databases. Also, the Canadian roadside sample program4 may inform the selection of the vehicles to sample for data collection. However, none of the current technologies can currently provide reliable estimates of ton-miles and value-miles of international trade traffic. Furthermore, characteristics of trade, especially in origins and destinations, are unlikely to be addressed by technological advances in passive information collection. In order to take advantage of many of these emerging technologies, an up-to-date, accurate geographic information system (GIS) coverage will be needed. Many states and local governments are using their own local GIS to locate and perform maintenance on their intelligent transportation systems equipment. Electronic maps with the geocoded location of these infrastructures would need to be “fused” across county and state boundaries to build a national network. The underlying infrastructures (including the devices producing the data streams) are dynamic and are intended to grow over time, which would require a dedicated, national effort to ensure accuracy and timeliness of the locations of these data generators. Mobile data sources (e.g., global positioning system [GPS] units on trucks) also need to be identified spatially. Without advances in the ability to upgrade and maintain the necessary GIS data, many of the potential data sources cited in this report will remain underused. The rest of this section briefly discusses eight emerging technologies: 4   The Canadian roadside sample program is an intercept survey of commercial trucks at U.S.-Canada border crossings.

OCR for page 20
Measuring International Trade on U.S. Highways weigh-in-motion; virtual weigh station; electronic screening, including radio frequency identification and dedicated short-range communications; smart trucks, which use global positioning and similar systems; bar codes; video image detection; vehicle and card inspection; and archived data user service. For each technology, each description is followed by comments on its advantages and disadvantages, cost, and likelihood of deployment. Weigh-In-Motion Technology Weigh-in-motion (WIM) technology is the process of measuring the weights of axles on trucks while in motion. Various types of WIM technology are capable of highly accurate measurements for a particular speed. The more expensive WIM systems can accurately measure weights at “highway” speeds (± 10 percent accuracy for 95 percent of the trucks), while less expensive systems are accurate at slower speeds (± 15 percent accuracy for 95 percent of the trucks). Advantages and Disadvantages. The advantage of WIM technology is that it will allow the capture of fairly accurate truck weights, and it requires no personnel to operate. The disadvantage of WIM is that not every road is outfitted with WIM technology. Also, WIM can only provide weights—it cannot provide any commodity information that can be used for estimating value. Cost. The costs for WIM technology depend on the type of technology used. The initial equipment cost for piezoelectric technology can be as little as $2,500/lane, and the cost of single load cell can be as high as $39,000/lane.5 While initial costs vary greatly, the overall life-cycle costs are much more equivalent.6 5   International Road Dynamics, Inc. (2001). 6   Recent experience by CalTrans suggests that the current cost estimates of WIM might be too low, since full-speed WIM sensors require a huge investment in high-quality pavement to ensure that the truck does not bounce over the sensors. Similar unexpected cost considerations might apply to other estimates presented here, given the uncertainties that accompany new technology.

OCR for page 20
Measuring International Trade on U.S. Highways Likelihood for Deployment. Near Term: many jurisdictions are deploying WIM technology today and are seeking funding to add more WIM sites with plans for enforcement purposes. Virtual Weigh Station (VWS) Technology Virtual weigh station technology uses WIM, closed circuit television cameras, and, in some cases, optical character recognition technology (for license plate numbers). VWS allows jurisdictions to identify trucks, collect their weight data, identification data, and also their speed. The system is also used to capture data on overweight trucks. Advantages and Disadvantages. The advantage of VWS technology is that it can not only capture weight data, it can also capture speeds, and identify the trucking company and the specific vehicle (by license plate). Another advantage of VWS technology over a WIM-only site is that the image of the truck can indicate to the jurisdiction the commodity given sighting of, say, a logging truck, a car hauler, a tanker truck, or some other obvious kind of carrier. The disadvantage of this technology is that there is no information on the contents of a closed trailer, that is, the value—only the weight. Cost. The cost for VWS technology is approximately $150,000-$200,000 per site. This includes WIM technology, cameras, and computer storage equipment (and possibly a method for communicating the data to an enforcement vehicle or a weigh station site). Likelihood for Deployment. Near to mid-term: several jurisdictions are deploying VWS sites today; many others have VWS in their jurisdiction’s technology deployment plans. Electronic Screening Radio frequency identification (RFID) tags or dedicated short-range communication (DSRC) systems use radio frequency to communicate between a truck-mounted tag and a roadside tag reader. This communication link has many applications, one of which is to identify freight as it is loaded and offloaded from trucks. The unique ID number of the tag corresponds to information contained in a database that houses detailed information on the freight with which the tag is associated. This two-way communication link is used primarily on the interstate highway system to identify a specific truck (or company), run a quick database check to deter-

OCR for page 20
Measuring International Trade on U.S. Highways mine if the truck is in compliance (paid taxes, etc.), and it can also alert a driver to pull into a weigh station or port-of-entry to be inspected. Advantages and Disadvantages. The advantage of using RFID technology is that it can identify the motor carrier as the truck travels down the highway. If a jurisdiction has a database link to the truck and its manifest (most do not have this capability), a great deal of information can be obtained. The disadvantage of DSRC in this application is that it can only directly identify the carrier and the truck. There is no commodity information, nor is there any weight information for many DSRC sites. Another disadvantage is that only a very small percentage of trucks carry DSRC tags and participate in the electronic screening programs. Moreover, the various programs that do exist are not interoperable so a tag that can be identified in one state may not be recognized in another state, and the data are not a public record that can be used to track commodity flows by public agencies. Cost. The cost for DSRC screening technology is approximately $25-$75 (personal communication) per tag. Depending on the program the carrier enrolls with which the carrier may or may not have to purchase the tag. Likelihood for Deployment. Near term and ongoing: several jurisdictions are deploying electronic screening programs across the country, but participation is low. Smart Trucks Smart trucks refers to trucks that are equipped with such technologies as GPS, satellite communications, and on-board diagnostic and monitoring capabilities. These trucks have the capability of monitoring, reporting, and performing automated functions while the truck is in motion. For example, sensors may indicate that the trailer doors have been opened and send a message by satellite, or sensors may weigh the contents of the trailer and transmit this information to dispatch for billing purposes. These systems also have the capability to monitor engine performance: for example, they can allow monitors in central locations to conduct such operations as adjusting the air-fuel ratio while the truck is in motion without the driver’s knowledge; they can even disable the vehicle in case of theft. These systems can also locate a vehicle and determine if a truck has detoured from its scheduled route. Advantages and Disadvantages. The advantage of having smart trucks is constant communication with a motor carrier’s assets. The cargo

OCR for page 20
Measuring International Trade on U.S. Highways can be monitored for temperature, the engine’s performance can be observed, and vehicle speeds and braking applications can be monitored and reported. Also, the driver’s logs can be maintained electronically. The disadvantage, in the context of this report, is that this information is private: unless the motor carrier volunteers this information, a government jurisdiction will have difficulty obtaining it. And as with other systems, none of the information, operating conditions, location and freight security and temperature monitoring provides value information for the cargo. Cost. The costs for GPS tracking systems can start as low as $1,100 for commercial fleet tracking applications, but this does not include software.7 There are both less expensive and much more expensive options for tracking vehicles. Many of the features discussed above involve the integration of various technologies linked to satellite communications systems. The cost for this technology varies depending upon the complexity of the system. A system with pallet readers at two locations will be considerably cheaper than systems with loading and unloading facilities in every state across the nation. A typical tag for this application can cost as little as $1.00 per tag.8 Likelihood for Deployment. Near term and ongoing: several of the larger trucking companies are deploying this technology today; others are waiting for the benefits to outweigh the costs. Legislation is being developed that will allow tax incentives for the purchase of such technologies that can increase safety and security. Bar Codes Bar code readers are a common technology in freight tracking and many other industries. This communication link has many applications, one of which is to identify freight as it is loaded and offloaded from trucks. The unique bar code assigned to a package corresponds to information contained in a database that houses detailed information on the freight with which the tag is associated. 7   See Terratracker (2004) available: http://terratracker.com/html/fleet.html [Accessed August 27, 2004]. 8   Doug Argie at Peak Direct, authorized distributor for Zebra Technologies, personal communication.

OCR for page 20
Measuring International Trade on U.S. Highways Advantages and Disadvantages. The advantage of using bar code technology is that it can provide detailed information for each parcel loaded on a truck and that data can follow the freight from pick-up to delivery. The disadvantage of this technology is that the information is not a public record that can be used to track commodity flows by public agencies. Cost. The cost for this technology is relatively inexpensive but depends on the complexity of the system: Hand-held units can range between $200 and $30009 depending upon the technology. However, with the wide-spread use of bar codes, the cost of this technology is substantially lower than a closed proprietary system. Likelihood for Deployment. Short term: This technology is widely used today by package delivery companies, grocery stores, etc. Video Image Detection Video image detection systems use machine vision technology to compile and analyze traffic data collected with closed circuit television (CCTV) systems. Video image detection can be used to monitor freeway conditions, capture speeds, count vehicles, and classify vehicles. Advantages and Disadvantages. One advantage of video image detection is that no additional equipment is required from the freight community. Also, the true counts of commercial vehicles can be obtained. In addition, limited commodity information can be collected such as logging trucks, car carriers, tanker trucks, etc. A disadvantage is that weights cannot be collected, and there is also no way of identifying the contents of closed trailers. Cost. The cost for this technology can be relatively high, with color CCTV cameras costing from $10,000 to $50,000 and annual maintenance costs ranging from $200 to $1,000. In addition, one needs a communications link, software systems, and algorithms for automated surveillance (Loukakos, online ).10 Likelihood for Deployment. Mid- to long-term: there are more than 5,000 video image detection systems now in operation, but most of these systems are outside of the United States. 9   System ID Warehouse: Available: http://www.systemid.com/BARCODE_SCANNERS/?source=overture [Accessed August 27, 2004]. 10   “Video-Image Detection” Dimitri Loukakos, UC Berkeley and Caltrans; Available: http://www.calccit.org/itsdecision/ [Accessed August 27, 2004].

OCR for page 20
Measuring International Trade on U.S. Highways Vehicle and Cargo Inspection System A vehicle and cargo inspection system (VACIS) technology uses gamma radiation to essentially take a “picture” of the contents of vehicles. This technology has located and identified drugs, people, and also identified cargo that was not listed in a cargo manifest. Advantages and Disadvantages. The advantage of VACIS technology is its ability to see the inside of vehicles without removing their contents. It has proven effective for security purposes in many locations for trucks as well as passenger vehicles. The disadvantage of this system is that inspection speeds are very slow, requiring trucks to stop; in practical terms, this means that not all trucks can be inspected. Cost. The cost for the VACIS machine is approximately $1M per unit.11 This technology is available for permanent installation locations as well as temporary setups used to inspect individual pallets of freight. Likelihood for Deployment. Mid-term: there are limited deployments of the VACIS technology today for roadside, ports, and railroads. However, in today’s security environment, more deployments in the near term are likely. Archived Data User Service Archived data user service (ADUS) is the storage of transportation-related data, for a given period of time, for use by anyone for planning purposes, trend identification, etc. These data are primarily collected through means of intelligent transportation systems. Many of the technologies mentioned in this report can support an archived data user service. Advantages and Disadvantages. An advantage of using ADUS is that there is a great deal of information collected today that is purged at set intervals. Through ADUS, these data could be collected and used for freight movement data, and possibly also for origin-destination data. The major disadvantage is that most of these data are lacking in specific information for a given vehicle. For example, video, WIM, or other technologies can only identify a truck—not which truck. Cost. The cost for ADUS is hard to quantify. Some systems already have ADUS capabilities built in, so the cost is nil. Other systems may 11   Brian McDaniel, Science Applications International Corporation, San Diego, personal communication.

OCR for page 20
Measuring International Trade on U.S. Highways require modifications to manage such large amounts of data—including scrubbing the data and supplying notes on data gaps or documenting when data collection systems are out of calibration or offline due to maintenance, etc. Likelihood for Deployment. Mid-term: although some ADUS systems are currently up and running, ADUS standards are still in development. Lately, more information technology procurements are including ADUS requirements.