A Methodology for Predicting Performances of Map-Matching Algorithms

Map matching is not always perfect and sometimes produces mismatches. Thus, there is a degree of uncertainty for how well a map-matching algorithm will perform under certain circumstances. Circumstantial factors include accuracies of sensor data and surrounding road network structure, among others. This paper attempts to shed light on this uncertainty and proposes a methodology for predicting performances of map matching algorithms at given locations on a digital road network. In short, using a vehicle’s position, the proposed methodology can be employed to predict the performance of a mapmatching algorithm at that position. Since map-matching algorithms are different in their logic of matching vehicle’s positions to road segments, there should be a separate prediction algorithm based on the methodology for each mapmatching algorithm. To demonstrate the methodology’s benefits, a probability algorithm to predict the performance of a point-to-curve map-matching algo-