Real-Time Context-Aware Computing with Applications in Civil Infrastructure Systems.

Abstract

This dissertation contributes a structured understanding of the fundamental processes involved in developing context-aware computing applications for the civil infrastructure industry. The civil infrastructure industry is characterized by mobile human and machine agents actively engaged in real-time decision-making tasks in a dynamic and unstructured workspace environment. This distinguishes context-aware computing from other computing technologies in three aspects: 1) it has the ability to perceive, interpret, and adapt to the agent’s evolving workspace; 2) It streamlines project data and presents the agent with information pertinent to its context, thus eliminating the agent’s tasks to accomplish the same; 3) By leveraging contextual information, it supplements decision-making tasks in real-time.

This research has successfully investigated technical approaches to address fundamental aspects of introducing context-aware applications to civil engineering, including: the ubiquitous localization of mobile agents in dynamic, unstructured environments; abstraction of the spatial-context and identifying the objects of interest to the agent; and the suitability of using standard models to manage and organize data for context-aware computing applications. A computational framework for designing context-aware applications to support real-time decision-making has also been implemented. The framework allows researchers and other end users to leverage currently available context-sensing technology to design and implement innovative solutions to domain specific problems.

The researched methods have been validated through several experiments conducted at the University of Michigan, the National Institute of Standards and Technology, and the Michigan Department of Transportation. These experiments have resulted in the implementation of several applications – to support real-life decision-making tasks – that not only serve to illustrate the usefulness of the framework, but also have significant social and economic implications. Among these applications are the controlled drilling system that warns drilling personnel when the drill bit tip is about to strike rebar or utility lines, thus helping preserve the structural integrity of concrete decks and preventing utility strike accidents; an automated fault detection system that diagnoses faulty components of an underperforming HVAC distribution network; and an innovative bridge inspection solution that supports condition assessment decision-making, thus introducing objectivity to visual condition assessment by providing concurrence with the Structural Health Monitoring data.