TinyAdapt: An adaptation framework for sensor networks
Daniel Minder, Marcus Handte, Pedro José Marrón · 2010
Although algorithms for wireless sensor networks are usually optimised for a specific set of user preferences under certain network conditions, if the conditions or preferences change during run-time, the intrinsic parameters of the algorithms have to be changed accordingly or the algorithms have to be replaced. However, choosing a suitable set of algorithms and parameters can be a tedious and difficult task to do manually. In this paper, we present TinyAdapt, a novel adaptation framework for sensor networks, that performs this selection autonomously guided by previous results from simulations, testbeds or real deployments, measured network conditions and high-level user preferences. We show that its use in sensor networks improves their performance significantly with minimal overhead and that it increases their flexibility under changing conditions.