Automatic generation of topological indoor maps for real-time map-based localization and tracking

Personal location information is regarded as the most important contextual information transmitted in ubiquitous systems. Many pedestrian indoor localization systems rely on map-matching to constrain sensor errors. The maps required for computer aided localization and tracking need to incorporate a semantic structure. Such maps are not readily available and therefore most groups working on localization solutions manually create the required maps for specific testing scenarios. To provide a solution for map generation on a larger scale, we have developed a map generation toolkit that parses standard CAD-plans, to automatically generate topological maps for indoor environments. We propose a heuristic parser that separates superfluous data from the information depicting semantic building entities, e.g. rooms and doors. In our experiments approximately 95% of all structures were detected successfully. After the extraction we transform the extracted building information into an object-based building model designed for the application of fast particle-filter-based map-matching algorithms. A performance test with a typical filter implementation demonstrates that the model is sufficiently optimized to achieve pedestrian tracking and localization in real-time.