Trajectory Data Reduction in Wireless Sensor Networks

This work addresses the problem of balancing the trade-off between the energy cost due to communication and the accuracy of the tracking-based trajectories’ detection and representation in Wireless Sensor Networks (WSNs) settings. We consider some of the approaches used by the Moving Objects Databases (MOD) and Computational Geometry (CG) communities, and we demonstrate that with appropriate adaptation, they can yield significant benefits in terms of energy savings and, consequently, lifetime of a given WSN. Towards that, we developed distributed variations of three approaches for spatio-temporal data reduction – two heuristics (Dead-Reckoning and the Douglas-Peuker algorithm), and a variant of a CG-based optimal algorithm for polyline reduction. In addition, we examine different policies for managing the buffer used by the individual tracking nodes for storing the partial trajectory data. Lastly, we investigated the potential benefits of combining the different data-reduction approaches into ”hybrid” ones during tracking of a particular object’s trajectory. Our experiments demonstrate that the proposed methodologies can significantly reduce the network-wide energy expenses due to communication and increase the network lifetime. Research supported by the NSF-CNS 0910952. Authors addresses: Oliviu Ghica, Goce Trajcevski, Peter Scheuermann and Fan Zhou, Dept. of EECS, Northwestern University, Evanston, Il 60208; email {ocg474, goce, peters, fanz }@eecs.northwestern.edu; Ouri Wolfson and Ugo Buy, Dept. of CS, University of Illinois at Chicago, 851 S. Morgan St., Chicago, Il 60607; email: {wolfson, buy }@cs.uic.edu; Dennis Vaccaro, Defense Systems Division, Northrop Grumman Corp., Rolling Meadows, Il 60008; email:[email protected]