Uncertainty and Sensitivity Analysis Methods for Assessment of GPS and GIS Integrated Application for Winter Highway Maintenance and Travel Time Study

The quality of spatial data has become an important issue since GIS began to be recognized as a practical tool to manage spatial information for a wide range of transportation applications. To maintain and update data for the applications, data collection methods often involve data collection vehicles equipped with DGPS receivers and numerous sensors. GPS and GIS are synergized when they are integrated into one application, but the capability to manage and analyze transportation data is reduced by positional uncertainties in GPS data and roadway spatial databases. Furthermore, the uncertainty is increased when associating two-dimensional (or three-dimensional) coordinates of GPS data with a one-dimensional roadway network model.;To deal with problems concerning the quality of input data as well as that of output information, this thesis presents sensitivity and uncertainty analysis methods developed based upon analytical and simulation error models for input data for the applications. Sensitivity analysis aims to quantify contributions of positional uncertainties in input data to variations in output information. Uncertainty analysis aims to estimate overall quality of output information derived from given positional quality of input data. However, due to the different approaches for modeling positional uncertainties and their propagation through spatial operations and computational models, the analytical and simulation methods for uncertainty analysis are validated and evaluated with spatial datasets selected concerning the curvilinearity and complexity of roadways.;The sensitivity and uncertainty analysis methods are applied to a winter maintenance application and a travel time study. The optimum input dataset for each application is determined by sensitivity analysis to output information computed from non-distance and distance based computation models. Results indicate that the winter maintenance application requires accurate input data since uncertainties in output information are accumulated as winter maintenance vehicles repeatedly treat roadways. However, for the travel time study, consistent output information is computed with a minute-accuracy-level, so positional uncertainties in input data have a negligible impact on output information.