Post-Disaster Road Traversability Mapping Based on GPS Track sharing and Map-Matching

Abstract

The main motivation in undertaking this research was to address some of the practical issues that were widely reported during major disaster situations such as the 11 March, 2011 Tohoku earthquake in Japan. Apart from great loss of life and property in the near vicinity of the earthquake, metropolitan areas in Tokyo experienced disruption of train services, electrical outages and inaccessibility to Internet services leaving many commuters stranded. It is, therefore, necessary to develop a near real-time traversability mapping service that can be available during or soon after the disaster. In this research, we implemented a system that facilitates generation of updated road traversability map using aggregated GPS tracks. The post-processing consists of a line simplification based on Douglas-Peucker algorithm and map-matching with existing road network using Hausdorff distance algorithm. Line simplification was applied reduce the data in further processing and map-matching to remove outliers and generate a navigable traversability road network. Performance of map-matching algorithm, evaluated using data collected in two field campaigns in Yamate-cho, Suita City and Sumiyoshi-ward, Osaka City, Japan clearly demonstrates the efficacy of the proposed workflow.