Third party funded individual grant
Start date : 01.05.2008
End date : 14.11.2013
Localization Systems (also known as Realtime Location Systems, or RTLS) become more and more popular in industry sectors such as logistics, automation and many more. These systems provide valueable information about whereabouts from objects at runtime. Therefore, processes can be traced, analyzed and optimized. Besides the research activities at the core of localization systems (like resilience and interference-free location technologies or methods for highly accurate positioning), there emerge algorithms and techniques to identify meaningful information for further processing steps. In this context, the aim of the wireless localization project is the research on automatic configuration methods for RTLSs as well as the generation of dynamic moving models and techniques for event processing on position streams at runtime. In 2009, we continued the development of our algorithms to estimate the receiving antenna's position (pose) of location systems. The algorithms estimate measuring points which allow a fast and accurate measurement pose. We applied an robot to execute the measurement automatically. The developed algorithm considers obstacles and the receiving characteristics of the location system and can sort out errors contained in the measurement data (e.g. multipath measurements). In 2010, models have been developed that can be used as dynamic moving models. Learning methods are applied to adapt the models at run-time. A formal language called TBL (Trajectory Behavior Language) was developed for describing trajectories. Additional algorithms can shrink the size of that description and hence compress the data size required to store trajectories. We are evaluating methods for learning the moving models online. These are evaluated in a study with respect to motion prediction. Moreover, it is being investigated whether events can be predicted by analyzing and learning event streams from the localisation system at runtime.