Grid Enabled Remote Instrumentation pp 81-95
Implementation of Average Consensus Protocols for Commercial Sensor Networks Platforms
In sensor networks, average consensus and gossiping algorithms, featuring only near neighbor communications, present advantages over flooding and epidemic algorithms in a number of distributed signal processing applications. This chapter looks into the implementation of average consensus algorithms within the constraints of current sensor network technology. Our event-based protocols work in the real event-based environment provided by a common Mica2 platform and use its wireless CSMA packet-switched network interface. Within this architecture our chapter derives different protocols according to an event-based software architecture that are suitable for an environment like TinyOS, the most used operating system for low-power mote platforms. Theoretical and simulation results are presented, and the main advantage over traditional routing protocols is given by the fully distributed and scalable nature this approach follows.
Unable to display preview. Download preview PDF.