Forming logical node groupings in sensor networks

Sensor network hardware has advanced to the point where cheap fully wireless nodes are available at commercially viable prices. Deployment of these nodes on any scale requires that they communicate with each other to aggregate and move data around the network. As many applications require the data to be combined from a particular set of nodes we propose a set of methods that make the logical groupings required by these applications possible. Our research focuses on the machine condition monitoring domain and how we can use the data generated by the sensor nodes to detect significant events in the time series data. We then describe a set of methods that can match these events and use the information to form logical groups based on the similarity of the events experienced. We then describe a simulation system that is being developed to investigate these methods along with the initial results and conclusions from our research so far