A distributed community based location service architecture

Location-based services (LBS) to date offer information and specific services tailored to the current geographic usage environment. The “location service” delivering the required location information in turn is provided by a location platform using specific positioning technologies and geodetic data (e.g. maps). Currently no single positioning technology or location platform can provide a location service as well as geodetic data for all known LBS usage environments and required location QoS (i.e. accuracy, delay, etc.). Hence scientific research agrees that a truly ubiquitous location service can only be obtained by use of heterogeneous positioning technologies in our daily living environment, that is, by use of dedicated systems such as GPS and technologies capable of also being used for positioning e.g. the abundance of heterogeneous networks all around us. The difficulty which remains is how to effectively “capture” this diversified technology data, generating meaningful location information and providing a homogeneous ubiquitous location service. In this thesis we present a novel distributed location service architecture which focuses on the collaborative generation and centralized provisioning of location information derived from heterogeneous networks and readily available positioning systems using distributed heterogeneous location platforms. The architecture is specified using an abstract system model. It contains functional components structured by three major core processes representing its initialization, positioning, and self-optimization properties. Client-based sensor mechanisms paired with centralized knowledge databases and community user models have been combined to provide a self-optimizing and domain independent location service framework. Within the service architecture a flexible Location Enabling Server (LES) framework has been developed, representing the initialization and sustainability processes of a LES for various dynamic usage environments. The initialization of a community-based LES with focus on collaborative, community-based network environment data acquisition for positioning purposes has been analyzed and successfully simulated for an outdoor innercity usage scenario. On the data processing side, suitable initialization and positioning models have been evaluated and empirically tested using captured WLAN hotspot information, showing the localization performance in relation to GPS within inner-cities. An important requirement for providing a ubiquitous location service is the discovery of distributed location platform entities with different geographic scopes of responsibility. Using the heterogeneous network and/or GPS sensor information discovered by the mobile client, we have developed a network independent, DHT-based peer-to-peer location platform lookup mechanism using XPath in a hierarchical Chord ring structure. A demonstrator has been implemented showing the effectiveness of this approach. Finally, a working location service architecture prototype has been implemented. As a proof of concept, a seamless indoor/outdoor navigation application has been implemented using two fully functional location platform entities.