A granular approach to aggregation within sensor networks
I.P.W. Sillitoe, J. Bergstrom · 2005
This paper presents a novel granular method of aggrega- tion. The method is suitable for use in simple resource-constrained sensor networks, in which full statistical estimation is not applica- ble. The method models the measurement uncertainty through the use of shadowed sets, and exploits the localised uncertainty within shad- owed sets to provide improved estimation plausibility. The paper de- scribes a general framework for the method and illustrates the advan- tages and flexibility of the approach with an example of the maximum value estimate problem. The results of the example are compared with those of a previous heuristic method.