Effects of Positioning Data Quality and Navigation Models on Map-Matching of Indoor Positioning Data

Abstract

With rising complexity of indoor environments and growing demand for positioning and tracking of people (such as occupants and field workers) indoors, there has been an increasing need to have accurate and reliable indoor positioning. In this paper, the effects of (1) the quality of positioning data and (2) the types of navigation models on the accuracy of map-matching of indoor positioning data are evaluated. Sensitivity analyses on the quality of two different types of positioning data, namely, (1) absolute point-positioning data and (2) relative point-positioning data have been carried out. Two different types of navigation models, namely, (1) network models and (2) metric models, have been evaluated to examine their effect on the accuracy of the map-matching results. Eight different map-matching algorithms were selected and their accuracies were assessed in six different indoor data-collection routes containing spaces with varying density, sizes, and shapes. The results show different ways through which the quality of positioning data and the navigation model type affect the performance of map-matching algorithms in different indoor layouts. In addition, no single algorithm generates the most accurate results for all the cases.