Abstract

As we observe a wider and wider deployment of Wireless Sensor Networks. In many cases, these systems are deployed in remote areas for environmental/wildlife monitoring purposes, and involve tens or even hundreds of sensors. They are expected to work without much human intervention, and for as long as possible. So far, the majority of these networks are not reusable or adaptive, in the sense that once the sensors are deployed, they cannot easily be reprogrammed unless someone collects them and connects them to a computer. This task is even more difficult (if not impossible) if the devices are attached to mobile entities or animals as these cannot easily be recollected. The current (few) solutions of the distribution of new code on a mobile sensor networks involved epidemic-like spreading of the code over the network. In this talklet I present a group-based code-dissemination protocol, which takes advantage of the dynamically built knowledge about the social relationships between the nodes to efficiently distribute the code; and a proposed system to selectively update specific subsets of the network.