In 1971, the National Bridge Inspection Standards (NBIS) (1) were promulgated in response to several catastrophic bridge collapses that could be traced to undiagnosed problems with these bridges. State and local governments began to systematically inventory the bridge population and collect vital condition information. The inspection and monitoring of bridges has thus continually evolved to the point where states and local governments are now implementing or adopting Bridge Management Systems (BMSs) to store and retrieve data on their bridge inventory and to support maintenance, repair, redesign and replacement decision making. States and municipalities have come to realize that in order to make sound infrastructure management decisions, they need to base their decisions on predictive models developed from accurate facility [bridge] condition data collected in the field (2). Effective bridge management is thus heavily dependent on field inspectors to collect detailed condition information on all of the individual elements of a bridge and enter this data into a bridge management system database.

To collect bridge condition data, the bridge inspectors must go to the bridge to assess its condition. Looking for damage and deterioration, they usually must: walk the deck and approaches to the bridge; climb on, under, and inside of the bridge superstructure; and examine the condition of foundations above and below the water surface. While advanced condition assessment technologies, such as video imaging, infrared thermography and ground—penetrating radar are being used for certain infrastructure applications, such as pavement inspection and bridge deck evaluation described in (3), visual inspection is still the primary means of data collection (4). Before going onto a bridge to conduct an inspection, inspectors prepare sketches and note templates to guide their inspection and to facilitate the recording of what they see in the field. Due to the nature of the inspection, the inspectors need to have their hands free to support their climbing and inspecting activities and cannot carry reference manuals, previous inspection reports, blueprints, coding manuals (i.e., manuals used to translate what the inspector sees into a specific coded level of damage) or other information with them onto the bridge. They only consult their manuals and the previous reports after leaving the bridge and returning to their inspection vehicles or to the office. Hence, during the inspection, the inspectors record what they see, then compare that information with blueprints, coding manuals, previous reports, etc. The inability of an inspector to access this supporting information during the inspection effort influences the efficiency and the quality of the collected field data.

Researchers at Carnegie Mellon have developed several generations of wearable computers specifically for inspection—oriented applications. The wearable computer provides users computing support in the field as they conduct other operations that require full use of their hands. In all of these applications, the emphasis has been on mobility, mostly hands free operation and either recording or displaying technical information. In one specific application, wearable computers were employed to assist U.S. Marines in conducting a limited technical inspection (LTI) of amphibious vehicles. This application of wearable computers lead to cost savings in terms of personnel needed (50% reduction), inspection time needed (40% reduction), and post—processing time needed for generating reports from the inspection data (30% reduction), for a total of 70% reduction in time from inspection through logistics data entry. In a second application, wearable computers were used to assist in the inspection of KC—135 aircraft, yielding a 50% reduction in data entry time. A third system supported maintenance of airport people movers providing remote access to over 120,000 engineering drawings, searching of previous maintenance reports, and telephonic and visual collaboration with remote personnel via a microphone and 3 oz. hand—held camera. In these applications, the reference material and inspection processes were the same for all instances of amphibious vehicles, KC—135 aircraft, and people movers respectively.

The use of wearable computers in bridge inspection has similarities to, and differences with, the previously described applications. The bridge inspection process, using a wearable computer, is similar in that the inspector is being supported in the field while doing an inspection. Bridge inspection is different from these previous applications in that each bridge, and the corresponding steps in its inspection process, may be unique. The necessary supporting information from each bridge will have to be collected ahead of time, as it is now, and entered into, or made accessible from, the wearable computer. In addition, different inspectors typically have different organizational schemes for this information. Some like to work from the actual as-built drawings, while some like to work off of abstracted framing plans. A second significant difference concerns the site. The previous inspection applications were performed in a maintenance facility. If a defect or error was found, it was tagged in some way and the defect remained tagged until a maintenance activity could be performed. Such tagging in the field is not feasible for bridges (with the exception of defects likely to lead to catastrophic failure, when an inspector can close or post a bridge immediately); the location and nature of damage must be accurately captured, recorded, and eventually entered into a bridge management system’s database. Maintenance decisions are then made by looking at the condition of the larger inventory of bridges; resource limitations may force some maintenance to be put off until a future time.

There are nearly 590,000 bridges in the National Bridge Inventory over 6.1 m (20 ft.) in length () that must be inspected at least every two years. There are many other bridges with spans less than 6.1 m (20 ft.) that agencies must still maintain at a cost to state taxpayers. Market research conducted by the Michael Baker Corporation (Baker) indicates that the annual cost of bridge inspections nationwide is approximately $400 million. While advanced bridge management systems are being created, deployed and used to collect this data for inventories of bridges and support the systematic identification and prioritization of needs, the bridge inspectors in the field are still primarily using paper—based notes to support, and record the results of, their inspection processes. Inspectors are currently limited in terms of the material they can reference while in the field and the tools they can use to capture the nature and location of defects found on bridge elements.

While most inspectors are still using pencil and paper on the site, some states and inspection agencies are experimenting with using computers on—site during the inspection process. All of the computing now provided on—site appears to be based on laptop, notebook and palmtop computers using a pen—based interface. For example, the Massachusetts Highway Department is using a system called IBIIS to store and manage all of their bridge documents (5). As part of this system, IBIIS provides inspectors with a "video camcorder and a notebook computer. The camcorder is used to take video and still photographs and the notebook computer is used to enter the NBIS rating data for each bridge into a database application and commentary into a word processor" (5). Another application is a system developed by the University of Central Florida for the Florida Department of Transportation (FDOT) (6). The system consists of both a field and office set up with a pen—based notebook computer used to collect all field inspection data.

The use of a notebook computer in the field is inconsistent with the bridge inspection process. Turner and Richardson (7) report that some states have experimented with having its inspectors enter inspection data directly into personal computers, but found that it is awkward to carry portable computers while walking around on a bridge. They also indicate that some experiments with small computers strapped to the inspectors arm or attached to a clipboard have been conducted by several states, but the "keyboard and viewing screen are so small that data entry can be very difficult" (7). Inspectors need to be able to use their hands for safe and effective inspection of a bridge and need to be able to view the data they are entering in less than optimal viewing conditions.

We are in the process of designing and implementing a prototype wearable computer to support bridge inspectors in the field as they inspect the individual elements of a bridge and record their findings. Wearable computers deal in information rather than programs, becoming tools in the user’s environment much like pencils or reference books. The wearable computer provides automatic, portable access to information. The information can be automatically accumulated by the system as the user interacts with and updates the information in the system, thereby eliminating the costly and error—prone process of information transferral. If we are able to achieve a savings in the inspection process similar to that seen in the amphibious vehicle inspection process (70% total time reduction), this could lead to significant annual savings that might be better employed in repairing and maintaining these bridges. Thus, there exists a significant potential for wearable computers to support bridge inspectors in the field that could lead to a decrease in cost and an improvement in quality of the inspections.

Lightweight portable computer systems now have the potential to support bridge inspection workers anytime and anyplace. Empirical field research has documented the value of high quality hypertext documents over paper, and mobile access to information. It is now technologically feasible to provide bridge inspectors with portable computers that are hardened to function where they may be dropped or hit and that have the capabilities and appropriate user interfaces to support faster and better bridge inspection.

Conventional palmtop computers, such as the Apple Newton and HP Omnibook, are clearly not suitable because they have limited connectivity to other computers. Some PDA products, such as Motorola’s Envoy, Sony’s Magic Link, and Bell South’s IBM—developed Simon, which depend on RF wide area networking (WAN) or wireless (typically IR) local area networking (LAN), also require two hands to operate.

To maximize the effectiveness of a wearable computer for supporting bridge inspection, the user interface design must be carefully matched with user tasks. By constructing mental models of user actions, interface elements may be chosen and tuned to meet the software and hardware requirements of specific procedures. The efficiency of the human—computer interaction is determined by the simplicity and clarity of the mental model suggested by the system. By modeling the actual task as well as the human interface, a linkage can be constructed between user and machine that can be examined to improve the overall efficiency of the wearable system (8).

Although the number of quantifiable metrics suited for interface evaluation is small, a series of basic observations provides a means for comparison. One characteristic of an application interface is the number of user actions required to perform given subtasks. We define a subtask as an operation, possibly consisting of multiple inputs, that a user completes in the process of performing a larger coherent task. For example, in the course of performing an inspection, a user might wish to return from their present location within an application to the main menu. This subtask may require a single input (perhaps a voice command or an on—screen button) or multiple inputs (backing out through a hierarchy of categories to reach the top, or main level). We assert that an application requiring few inputs will allow a user to dedicate more attention to the job at hand, while a lager number of inputs will require more concentration on the computing system. While a speech recognition engine accepts complex commands, which allows some subtasks requiring a series of manual inputs to be executed with a single phrase. However, the response time to a spoken input is longer and the accuracy is lower. For these reasons, we must factor in the quantitative aspect of system latency and accuracy into our evaluation of usability.

A second issue concerns what information should be provided, and in what form, to the inspectors via a wearable computer. We have been interviewing bridge inspectors to determine the information they need while in the process of an inspection. Based on these early interviews, we developed a prototype that we demonstrated to a different group of inspectors. These inspectors gave us a large amount of feedback usability. They recommended that the user interface be designed to make it easier to make selections of drawings and photos (we had small hotspots that were not visible until the mouse moved over the spot, which is a common functionality on web browsers). They asked for voice input and recognition. They do not want to type anything in the field. Inspectors must be able to quickly and easily identify the location of a defect and to record a textual description of it, take and store a photo of it, and possibly record an audio and/or video clip of the defect. In addition, inspectors must be able to sketch the defect as they often do now (even when taking a photo because often times the photos are not taken in the nest of conditions). It was clear from this initial prototype that these wearable computers for bridge inspectors must be developed from the beginning with consultation from bridge inspectors. The approach taken when developing all previous generations of wearable computers has been a user-centered design approach.

A third issue is that many inspections are conducted with multiple inspectors. Hence, a wearable computer must be designed to allow these inspectors to communicate and coordinate their actions while on the bridge. Currently, they use radios to coordinate themselves. A wearable computer would allow them to communicate not only words, but text, reports, images, drawings and sketches.

The bridges in the National Bridge Inventory must be inspected at least every two years. If wearable computers can be developed and used to support bridge inspectors, significant decreases in the cost of bridge inspections are possible. The quality of the inspections are also expected to improve due to the greater access by inspectors to information while they are in the field. However, unless the system is developed from the very beginning with continued participation and evaluation from the eventual users, it will not be actually used in the field.