Collaborative Spatial Decision-making in Cellular Network Management

Joseph DeLotto & Greg Theisen

Nynex Science and Technology, Inc.
Geographic Information Systems Group
Network Planning & Engineering Systems Lab
500 Westchester Ave., White Plains NY 10604

Introduction

Cellular communications is a highly fertile subject matter domain for the study of Collaborative Spatial Decision-making (CSDM) processes. The Cellular industry has both sophisticated spatial analysis requirements and, in general, large and complex organizational structures. These factors, combined with a highly developed computing and communications infrastructure, create an environment where CSDM can have a significant business impact. The goal of our work is to assess the potential benefit that techniques from collaborative work can contribute to our existing spatial analysis environment.

This paper will outline a task model for CSDM in Cellular network management, and present a software environment of existing spatial databases and associated applications that can serve as a starting point for the creation of `cel- lular groupware' for spatial decision-making. While the emphasis is on the problem domain, strategies to create group based spatial decision tools are sug gested as a result of our initial experience.

Spatial Decision-making in Cellular Communications

Where is the greatest demand for cellular service? How will radio signals behave in different geographies? Where should new technologies or services be deployed to maximize their cost-effectiveness? These are examples of the types of questions commonly encountered in the cellular industry. To answer these questions, telecommunications companies are making increasing use of GIS and spatial decision support systems (Ding et al, 1995). While these tools have contributed significantly to individual productivity, they have not been used to foster collaborative problem solving or increase organizational learning.

Almost every facet of building and operating a cellular network depends on a spatial decision process. Planners need to know where, when, and how much subscribers will use their phones so that they can determine how capacity should be distributed across a service area. Marketing departments need to know both where service is offered, and who within that service area is likely to be a potential subscriber. Engineers must model potential network configu- rations to insure that the appropriate signal quality is provided to areas identi- fied by planners. The illustration below is a generalized view of the functional organizations responsible for operating the network, and the primary flow of information among them.

Figure 1 - Organizational Structure

Obviously, the organization encompasses a broad range of spatial decision problems which no single GIS application can support. Individual departments, such as engineering, represent large-scale work environments in their own right. The engineering department at Nynex Mobile has over 200 engi- neers in a number of functional subgroups, located at multiple offices through- out the Northeast. All departments use some type of GIS-based tools in their analysis and reporting.

Individual departments are highly independent, and tend to collect a variety of spatial data on an ad hoc basis. This data is used to produce maps showing the distribution of some feature or the result of an analytical model. Maps are used as a static representation of a feature, to be shown in meetings or included in a report, rather than as a dynamic model of the decision space. The challenge for system designers is to allow decision-makers to make more effective use of spatial data both within and between departments. Within departments, the problem is similar to other large-scale technical project management studies (Grønbæk et al, 1992). This case presents a conventional notion of collaborative work where individuals contribute towards a larger group goal.

To address the use of spatial data between departments is to apply the notion of collaborative work to the exchange of information and ideas among multiple functional groups within an organization. From a social perspective it may be viewed as another case of individuals, in this case department managers, using collaborative methods to solve problems. From the perspective of designing tools that allow participants to exchange ideas in a spatial context, it presents a problem of creating generalized spatial representations of functions conducted by that department. The rest of this paper addresses the current methods used to facilitate interdepartment spatial decision-making.

An Example Using Existing Applications

To explore the collaborative use of spatial data and analysis for the activities outlined above, we have created a software environment from existing GIS applications. Two key objectives of the environment were to encourage the concept of a map as dynamic workspace, and to promote a common spatial database which allows easy exchange of spatial information among users. The architecture of the environment is shown in figure 2.

The core of the environment is the Signal Quality System (SQS), which provides the central database for spatial and network data. SQS is primarily an engineering tool, so it contains highly detailed data on network parameters and activity. SQS is based on a shared database which allows users to access data on the actual state of the network and create their own views of the network. These `cellular views' represent a snapshot of an actual or proposed network configuration, and are made up of both spatial and aspatial data. A cellular view has a multi-layer geographical representation composed of both raster and vector data which are linked to relational data about individual network components. They can be compared to other cellular views, or referenced by external applications for non-engineering purposes.

GIS applications in other departments, such as operations or regulatory, can access a common spatial database as well as cellular views created in SQS by various users. The object of collaboration is now the cellular view, and although the view cannot be edited by multiple users, other applications can derive new data or relate other attributes to it. The multi-layer view can replace a map as the media for exchange of information between departments.

Figure 2 - Spatial Data Analysis Infrastructure

Advantages and Disadvantages of Using Existing Applications

The environment described above is an attempt to introduce some concepts of groupware to a spatial analysis problem by building on existing GIS applications. This pragmatic approach satisfies the broad definition of groupware as "the computer-based systems that support groups of people engaged in a common task (or goal) and that provide an interface to a shared environment." (P. 40, Ellis et al., 1991). Even though it is based on existing computing infrastructure rather than some formal model of communication, it provides an initial platform to experiment with further system design.

We have tried to leverage the inherent ability of maps as a communication tool to provide a starting point for the development of CSDM support. However, simply providing a map interface does not promote collaborative work. Collaborative tools must promote greater interaction in actual work situations. We have attempted to do this by providing application specific user interfaces, a user-modifiable spatial modeling environment, and easy access to data and applications through a variety of communication protocols.

One obvious weakness in this approach is the lack of support for simultaneous interaction between workers at different sites. Spatial data may be created and viewed by a single user from any location, but there is no capability for a user to modify or suggest changes to another user's data set. A WYSIWIS map editing capability where all users interactively provide input would provide the ideal solution for this purpose. Existing applications also lack strong version- ing procedures to track various stages in the decision process. Multiple cellular views can be created and compared in user workspaces, but no meta-data about these are maintained by the applications. Lastly, the retrieval and processing of spatial data is still slow, and will have to be vastly improved to be useful for interactive decision making capability.

Future directions

There is clearly room to improve the way that spatial data and analysis are utilized in the design and operation of cellular networks. The component GIS functionality already exists or is readily available, but the systems are not designed to support group use. Before further development can occur several issues need to be explored, including:

Collaborative methods will become even more necessary as cellular operators establish national footprints over the next several years and support organizations are distributed geographically. GIS applications are a vital component in network management, and will have to provide some level of groupware functionality.

Bibliography

Ding, Y., Boguchwal, L., DeLotto, J., Fischman, G., and Theisen, G., "GIS Supports the Design, Planning and Engineering of Wireless Communication Networks" Proceedings forthcoming, Geoinformatics '95. Hong Kong. May 25-28

Ellis, C.A., Gibbs, S.J., and Rein, G.L. "Groupware: Some Issues and Experiences", Communications of the ACM, Vol. 34, No. 1, January, 1991. pp. 38-58

Grønbæk, K., Kyng, M., and Mogensen, P. "CSCW challenges in large-scale technical projects - a case study", Proceedings ACM 1992 Conference on Computer-Supported Cooperative Work (1992), ACM Press, pp. 338-345

Biography

Joseph DeLotto is a Member of Technical Staff in the Planning and Engineering Laboratory at Nynex Science and Technology. His research interests are in radio propagation modelling, spatial data models, and the geography of telecommunications.

Greg Theisen is a Member of Technical Staff at Nextwave Communications. He previously worked in the Planning and Engineering Laboratory at Nynex Science and Technology. He has worked on several wireless planning projects, including the Signal Quality System (SQS), a GIS based planning and engineering tool. Current research interests are cellular demand forecasting and wireless technologies.