Historic Queries in Geosensor Networks Stephan Winter 1 ,S¨ oren Dupke 12 , Lin Jie Guan 1 , and Carlos Vieira 13 1 Department of Geomatics, The University of Melbourne, Victoria 3010, Australia 2 Institute of Geoinformatics, University of M¨ unster, Germany 3 Departamento de Engenharia Civil, Universidade Federal de Vicosa, Brasil winter@unimelb.edu.au Abstract. This paper addresses—to our knowledge, for the first time—the prob- lem of querying a geosensor network—a sensor network of mobile, location- aware nodes—for historical data. It compares different network architectures and querying strategies with respect to their performance in reconstructing events or processes that happened in the past, trying to support the hypothesis that recon- struction is possible within the limited capacities of a geosensor network only. In a concrete case study, these queries are studied in a simulated peer-to-peer ride sharing system. 1 Introduction Mining for historic data is relevant to study complex systems. Just querying the current states and plans of self-organizing, mobile and autonomous agents may never show the higher patterns of behavior emerging from their activities. The running example in this paper will be an ad-hoc peer-to-peer system providing shared rides for partici- pating pedestrians (clients) and vehicles (hosts) in urban traffic [1–3]. In this system, agents negotiate shared rides with local peers, move, and regularly update their travel arrangements according to new opportunities emerging in their current neighborhood. This means, a client’s complete trip can only be queried in hindsight, after completion. Considering recent trends in ubiquitous computing and ambient intelligence, com- plex systems in geographic scales and environments will be observed, managed and analyzed by intelligent computing platforms that are sensor-enabled, mobile or embed- ded, and wirelessly connected to peers or web-based services. This connection of the intelligent computing platforms can be based on different architectures. Location-based services, as one example, connect the mobile device of an individual agent in a complex system directly with a central service via mobile communication operators. Intelligent transportation systems, as another example, typically have a hierarchical structure of localized, regional and centralized sensor analysis and traffic management operations. Finally, peer-to-peer services such as the mentioned shared ride system, or any other geosensor network, work by local collaboration and have by design no need for a cen- tralized service. This paper addresses the problem of querying the mobile intelligent computing plat- forms for historical information. While this problem is trivial for data collected in a central or even in a federated database, it is not trivial for data circulating in geosen- sor networks. In geosensor networks, agents can, for relatively cheap costs, temporarily This is not the authoritative version, which is only available from Springer in Ware, M.; Taylor, G. (Eds.), 7th International Symposium on Web and Wireless GIS.