GISdevelopment.net ---> GITA 1998 ---> Field Applications

The Digital Pen: Mapping and Data Acquisition

Kerry Smyth
Pacific Bell
2700 Watt Avenue, Room 4022
Sacramento, CA 95821


Context
The context for this paper is based upon project experience derived from the Consumer Broadband Initiative launched by Pacific Bell in November 1993. The $16 billion dollar broadband project called for deployment of a new hybrid fiber-coax network which would support voice, high speed data, and video services over a single wire feed. Two key factors of the project created the opportunity and necessity to deploy the most current AM/FM field technology available in the market place.
  1. The business case dictated expense savings by creating an intelligent network to reduce operation and maintenance expense. Therefore, advanced system applications were mandatory to create the intelligence of the network.
  2. The broadband hybrid fiber-coax architecture required an accurate landbase and facility mapping application to properly design and deploy the network.
It is important to understand the characteristics of the fiber-coax architecture and the nature of “Radio Frequency Design” (RFDesign) to appreciate the critical role the design and mapping systems would play in creating the new broadband network. In very basic terms, every phone is connected to the telephone company’s central office (CO) by a pair of twisted copper wires. If that CO serves 20,000 phones, then a network of 20,000 individual copper wires span the geography served by the CO. With the hybrid fiber-coax architecture, fiber-optic cable extends to the various regions and neighborhoods served by the CO. The specifications of the network architecture dictate how far the fiber extends to the consumer, and when the transport medium switches from fiber to coax for final distribution to the consumer. The Pacific Bell architecture called for the fiber to extend into a CO service region called a node boundary. which would serve approx !imatel y 500 residences.

The mapping challenge was therefore to define accurate node boundaries, with accurate address and service line requirements so the network could be designed for maximum efficiency in both deployment and long term operational efficiency. To accomplish this task, an accurate landbase was obviously necessary to get accurate service line counts within each node. And, an accurate field inventory of existing plant (poles, conduits, etc.) was required so the new network design path could be routed to “piggy back” the existing plant where possible.

System Development Strategies
It was clear that systems development and integration would play a critical role in the design, deployment and operation of the new broadband network. The question became what guiding principle would produce the systems and processes necessary for the project? The answer was simple and direct: Data. Data capture and data flow through the entire systems architecture was critical to creating the intelligent network. System development strategies to support the data requirements of creating the intelligent network included:
  • Structured Data Capture: Treat each piece of information gathered or created as a data element. If the information was worth capturing, apply the rigors of data management to increase its quality.
  • Capture at the source: Direct capture of information into digital data eliminates transcription and the system applications by design will produce “structured data.”
  • Data independence from graphics: Utilize systems and applications which support open or transportable data to assure data flow-through with other systems.
  • Reuse of Data: If the above strategies were applied, then data should be reusable through many downstream applications rather than having to be re-created, thereby introducing duplication, discrepancy, etc.
Given these strategies for system design, how could they be applied to applications which spanned GIS. CAD, AM/FM and database technologies? And more importantly, provide a system solution where over half of the work effort occurs in the field? Specifically, the system architecture had to support and enable the business to:
  • Deploy hundreds of automated field technicians?
  • Draw thousands of maps required for design and construction?
  • Design tens of thousands of nodes?
  • Validate millions of addresses?
  • Inventory tens of millions of poles and plant elements?
The Digital Pen
The digital pen enabled broadband field technicians to digitally capture, edit and validate data in the field. By doing so, the system strategies were realized (structured data capture, capture at the source, data independence, re-usability) and the spectrum of product deliverables from maps to address databases were delivered.

The digital pen was deployed at Pacific Bell via the Field Data Unit (FDU). The FDU was a 486 40 Mhz “pen slate” computer (state of the art in 1995) with a pen-based application environment. The pen slate computer provided a 9. 5“ monochrome diagonal screen for the display of digital maps. The unit was entirely self contained and all user interaction was via the digital pen. The pen application environment provided a variety of interface options including, on-screen digital keyboards, pick lists, radio buttons, data forms, and graphic features and functions to support field data capture. The detailed resolution provided by the integrated digitizer, display, and pen accuracy proved critical to the broadband mapping application.

The key requirement for any pen based application is the support and integration of the different data types to be captured and/or verified in the field. At Pacific, this required an application environment which supported the integration of landbase parcel maps, address databases, and graphic symbols for outside plant inventory and their associated data attributes. Typically, these data types span multiple systems and this was true at Pacific as the FDU pen application supported data flow to and from both the GIS and CAD systems used in designing the network.

The digital pen was not used to draw every map or design every node boundary. However, the ability to capture, verify and integrate data from multiple systems and the field proved to be a key role of the FDU in the total broadband systems architecture.

Additional detail regarding the specific system utilized at Pacific Bell is documented in a series of three articles published over the three year life span of the project. Please see the References section.

The remainder of the paper will focus on critical success factors, key learning experiences, and return on investment to position future users for success with the digital pen in field automation and mapping.

Critical Success Factors

Management Support
No surprise here! The data collected by a digital pen application usually serves a variety of client organizations. Therefore, top management support is important to gain departmental support to implement cross organizational data flow. This is particularly important to realize downstream benefits.

Example: The Pacific FDU project was very unique in it had top management support from day one. The business case dictated an intelligent system, and that was impossible to create from paper maps and inventories - using the digital pen to produce digital data was the only option.

Financial Backing
Prototyping on a shoestring budget is possible. Full scale development and deployment of a production ready system on a shoestring budget is a recipe for failure. Today’s development tools can produce deceiving results. A pen application is not a stand-alone application - it requires integration with existing systems to create the data flow to and from the field for success. Be sure the complete financial framework of the project and support systems are studied - not just the development of one or two application forms on the pen computer.

Example: The FDU prototype application was up and running within four months of concept. The first “production release” did not occur for another eight months. This was primarily because the production value and functionality proved so important that the application literally doubled in sophistication before releasing to the field. And, doubled again within the next six months because of additional needs and benefits.

Dedicated -0 ualified Resources
Application development should include personnel experienced with the full systems development life cycle. Experienced project managers and developers position the project for success, not just deliver a collection of code. Training, deployment strategies, and production processes are just as important as the application itself.

Example: The dynamics of designing a field application for a network with changing technical specifications required a flexible development team. A majority of the application was specified over the phone and e-mail. This type of relationship was only possible because of the professional experience of the programming team.

Field Participation
Field participation in specifying, developing and testing a pen based system is critical. Get the field users involved early. Identify individuals that are willing to take the “digital” leap and be sure management supports dedicated participation by selected field engineers. And most importantly, develop product champions early - do whatever it takes to make the first users successful.

Example: The first team trained at Pacific was called the “Gold” team. They committed themselves to making the FDU a success. As the FDU was deployed more broadly, it was received with various levels of enthusiasm. However, once they saw first-hand the productivity of the “Gold” team - they never questioned the FDU again.

Time
Give the project time to succeed. Good projects will fail given unrealistic time constraints. Design the project for incremental success, not an all or nothing deadline. This will allow the project to deliver successful milestones. Incremental milestones provide flexibility to alter course based upon changing business conditions and retain support during the process. Example: The FDU project was given time to succeed. However, managing expectations and delivery schedules can be very difficult when the end users are actively involved in the development cycle. Once something is seen, even if only a prototype - the expectation is it will be available next week. This is a good problem to have, but it is a double edged sword. There is only one first impression - don’t jeopardize the entire project with an unstable first release.

Key Learning Experiences
The automated field system is only as strong as its weakest link. If the battery fails, the entire system fails. If the screen is unreadable, the entire system fails. The total system is the solution - not the pen type, not the battery type, not the display type.

Field test! Throw away the specification sheets. Make product decisions in the field, not in the office.

Field Test! Use real data sets. Conduct performance testing with real data sets of the same size and complexity as the production data sets.

Understand Your user. Understand what is required and why it is wanted or perceived to be required. Cross check application requests and enhancements against magnitude of impact on total production process. Invest and develop application functionality serving the dominating production tasks, not the secondary ones - even if they are not as flashy.

Don’t short change training. Develop training curriculum which includes all necessary instruction and includes field training without production responsibilities. Position the users for success. Survey their computer literacy. Implement pre-requishe “basic” computer classes if necessary.

If possible, include an experienced user in the classroom to supplement the professional trainer with first-hand field experience and orientation for the new users. Provide on-site, walk-along field support during deployment.

Retuen on Investment
The original FDU business case written in 1995 called for approximately 25’XOsavings in fielding time based upon preliminary field trials. The basic premise was to eliminate the duplication of effort required to hand draw field maps and then re-draw the same information on the CAD system. Because drawing transcription would be eliminated, quality check points and re-fielding could be eliminated. This simple and conservative model produced an acceptable ROI.

This model represents a “sub-optimized” ROI calculation. The ROI was calculated only on the tasks that were directly automated by the application. Downstream benefits were known to exist, but were considered intangible, difficult to quantify, and out of scope because they relied on other systems so they were not included in the business case.

It is precisely these “intangibles” that proved to be the most important benefits of the FDU project. These benefits were in the “orders of magnitude” not just 25%. Although only a guesstimate, savings and benefits could probably be calculated in the 300-500% range or more.

Benefits
The conventional process of marking up paper maps in the field proved inaccurate and incomplete, therefore numerous re-walks were required. With the FDU, the software-aided process produced a more complete and accurate survey eliminating re-walks. Every re-walk eliminated represented an order of magnitude in savings.

The FDU deliverable was digital and supported color plotting. During a FDU field validation, the ability to digitally “redline” existing maps in color provided an excellent audit trial for separating the original condition from the new field edits. If color markings were applied to paper maps, the color was lost if a copier was used for distribution in the office. The quality and clarity of the color digital plot directly from the field was particularly valuable to node designers. The designer could easily identify the “redline” edits and quickly see if the field edit impacted the design by plotting the “redlines” over the existing network design.

The address database files that were collected and verified on the FDU transferred directly into the GIS system, eliminating tedious data input. GIS data bashing routines were then used to produce landbase accuracy’s exceeding 990A. The benefits of accurate landbase addresses correlated directly to identifying potential customers and capturing market share - significant to any business. In this case, if existing landbase was assumed 80°/0 accurate upon acquisition, the FDU and GIS data processing systems increased the market potential by 25%.

An accurate and electronic landbase supported automated node bounding. Automated node bounding and numerous design alternatives could be accomplished in seconds, whereas turnaround time for manual node bounding and the corresponding address list production took days. Again, resulting in orders of magnitude in savings. This is a classic example of a downstream benefit from an upstream automation effort.

As the field, mapping, design and engineering systems came on line and were integrated, the benefits over manual or stand-alone automation efforts validated the benefits of the original system development strategies.

Conclusion
The digital pen produces increased efficiencies and higher quality work products in the field. More importantly, the pen’s digital information opens entirely new opportunities for a business to streamline and automate their mainstream work processes because the field data is now digital.

References
  • Curry, S. and Jimenez, J., September 1995, Dial Tone of The Future - Mapping the Fast Lane to the Information Superhighway, Geo Info Systems
  • McFall, K., June 1996, Analyzing Pacific Bell’s Automation Opportunities, Geo Info Systems Wilson, J., March 1997, Field Computing Strikes Gold at Pacific Bell, Earth Observation Magazine
© GISdevelopment.net. All rights reserved.