Contributions to Distributed Spatial Compression in Wireless Sensor Networks
Javier Pérez Trufero · 2010
This thesis presents several contributions in the field of distributed spatial compression in Wireless Sensor Networks. First, since in most of the spatial compression schemes some nodes (raw nodes) need to broadcast their raw data to allow other nodes (aggregating nodes) to perform compression, we design several distributed heuristics which, via local communications, split the nodes into raw/aggregating subsets and optimize the amount of energy consumed in the network. We also extend previous work in the use of graph-based lifting transforms for data compression in distributed data gathering applications, to networks with more than one sink, and scenarios where all data has to be available at every node. Additionally, under the scope of these contributions, we design a new energy-efficient multicast routing algorithm, which is based on the minimum Steiner tree and exploits the broadcast property of wireless communications. We prove via computer-based simulations that our methods reduce the energy consumption in the network in comparison with existing approaches.