Transportation linear referencing toolboxes : a 'reflective practitioner's' design approach

Abstract

Seventy percent of the data of a typical transportation agency (e.g., bridges, accidents, etc.) has location as a primary reference. A Linear Referencing System (LRS) is the main way of identifying the location of this data and providing a storage key for it in a database. LRS is based on a one-dimensional offset on a predefined network. In theory, it is one of the simplest spatial cases. In reality, it can be spatially and analytically quite complex. LRS to quite recent date has been little formally researched. That research which has occurred has been the construction of large and comprehensive conceptual data models. This thesis is not primarily aimed at new "tool building research". The existing models have been based to only a limited extent on a fuller analysis of the nature of transportation and spatial data; they have not considered relevant field and wider methodological concerns (i.e., they followed a "model-driven" approach). The goal here is to create a more appropriate foundation and base from which LRS tools may be most appropriately built (i.e., a 'field-driven" approach). A "practitioners perspective" view of LRS was sought. Such a more holistic understanding was sought through the adoption of a "layered methodology" of research that involved gaining the perspectives of a variety of disciplinary viewpoints. This research framework was developed especially for this thesis based on the ideas and work of Schon and Reich. The approach involved in short a desk exercise in fundamental consideration of the nature of LRS, a deeper, cross-field synthesis and literature research, four in-depth state DOT LRS case studies, a panel of transportation field experts, a panel of national data model experts, and a limited object-orientated modeling exercise. The conclusion reached is that while LRS in the simple case can be modeled in general forms, it is also an "exception-driven" field. Thus, a "toolkit approach" may be more appropriate for LRS. It is inferred that this may hold for other similar application areas in transportation and planning. Further research would further develop the holistic layered methodology adopted here and further define the proposed LRS transportation application toolboxes.