An interesting and novel algorithm called SEAD for the construction and maintenance of a minimum spanning Steiner tree, weighted by energy consumption (d-trees), is presented in this paper. Its main goal is to minimize communication energy between fixed sensor sources and one or more mobile sinks. The algorithm is to be used in scenarios where robots or humans carrying personal digital assistants (PDAs) move around in an area where sensor motes have previously been dispatched and organized around replicators and access nodes. The PDAs (the sinks) collect live sensed information about noise, air quality, contamination, and so on.
A detailed evaluation of the algorithm relies on the MICA2 mote model, with the TinyOS Java operating system, and mobility simulation with Network Simulator-2. Main results show that SEAD consumes less energy per node when compared with similar algorithms, namely, directed diffusion (DD), two-tier data dissemination (TTDD), and the Internet Engineering Task Force (IETF) adaptive demand-driven multicast routing (ADMR). Another evaluation analyzes the impact of sink mobility on the performance of the SEAD algorithm under various patterns (such as speed and the waypoint model), and indicates that with SEAD, more energy per node is saved than with TTDD or DD, but that energy distribution among nodes is better for the latter two algorithms.
A full section is dedicated to a review of related work, and offers a rough feature comparison between SEAD and other energy-oriented sensor networks methods, such as DD, SAFE, TTDD, and ShopParent. Further Internet-related approaches, such as geocasting and the results of the OceanStore project, are also briefly mentioned in the comparison.