Distributed Algorithms for Location Based Services

Abstract

Real-time localization services are some of the most challenging and interesting mobile broadband applications in the Location Based Services (LBS) world. They are gaining more and more importance for a broad range of applications, such as road/highway monitoring, emergency management, social networking, and advertising. This Ph.D. thesis focuses on the problem of defining a new category of decentralized peer-to-peer (P2P) algorithms for LBS. We aim at defining a P2P overlay where each participant can efficiently retrieve node and resource information (data or services) located near any chosen geographic position. The idea is that the responsibility and the required resources for maintaining information about position of active users are properly distributed among nodes, for which a change in the set of participants causes only a minimal amount of disruption without reducing the quality of provided services. In this thesis we will assess the validity of the proposed model through a formal analysis of the routing protocol and a detailed simulative investigation of the designed overlay. We will depict a complete picture of involved parameters, how they affect the performance and how they can be configured to adapt the protocol to the requirements of several location based applications. Furthermore we will present two application scenarios (a smartphone based Traffic Information System and a large information management system for a SmartCity) where the designed protocol has been simulated and evaluated, as well as the first prototype of a real implementation of the overlay using both traditional PC nodes and Android mobile devices.