The Global Positioning System (GPS) and geographic information systems (GIS) have been integrated for a wide range of transportation applications. When a linear datum is not implemented into the applications, a route measure is indirectly determined with snapped GPS-derived coordinates along roadway centerlines. Thus, uncertainties in output information are subject to various types of error and error sources in GPS and roadway centerline maps and their propagation through spatial operations. A method for uncertainty analysis is presented. The method estimates the quality of output information from GPS- and GIS-integrated applications for transportation. The method involves an analytical GPS error model and an error propagation model, assuming that a test roadway centerline map is representative of roadway centerline maps with the same nominal scale. In the case study, the method was applied to a winter maintenance application and a travel time study. The optimum input data set for each application was determined by sensitivity analysis that explored the impact of positional uncertainties on variations of output information computed from distance-based and non-distance-based computation models. Results indicated that the winter maintenance application required accurate input data because uncertainties in output information were accumulated as winter maintenance vehicles repeatedly treated roadways. However, for the travel time study, consistent output information was computed with a minute level of accuracy, so positional uncertainties in input data had a negligible impact on output information.
Read full abstract