On the comparability of indoor localization systems' accuracy

Abstract

While indoor localization is gaining increased relevance, the comparability of the published accuracies for localization algorithms is rarely given since the metrics, node densities and the evaluation environments vary. To overcome this limitation, a simulator for indoor localization algorithms called simLoc is presented which uses prerecorded received signal strengths of two different buildings. simLoc simulates both, different path loss models and different localization algorithms. Therefore, the simulator can be used to compare different algorithms, but it can also be used to find optimal parameter settings. This was done to optimize a localization system for mobile devices equipped with a 868 MHz radio. As a result, the optimized algorithm achieves significantly lower error-distances, floor-errors, and localization-errors than the model-based RADAR system.