An efficient learning technique to predict link quality in WSN

Dana Marinca (1) , Pascale Minet (2) , Nesrine Ben Hassine (1, 2)

1 PRISM - Parallélisme, Réseaux, Systèmes, Modélisation
2 HIPERCOM2 - High PERformance COMmunications

In this paper, we apply learning techniques to predict link quality evolution in a wireless sensor network (WSN) and take advantage of wireless links with the best possible quality to improve the packet delivery rate. We model this problem as a forecaster prediction game based on the advice of several experts. The forecaster learns on-line how to adjust its prediction to better fit the environment metric values. Simulations using traces collected in a real WSN show the improvement of the prediction when the experts use the SES prediction strategy, whereas the forecaster uses the EWA learning strategy.

Mots clés: prediction; expert; cumulated loss; Machine learning; forecaster; link quality; wireless sensor networks

Domaines: Réseaux et télécommunications [cs.NI]

Dana Marinca, Pascale Minet, Nesrine Ben Hassine. An efficient learning technique to predict link quality in WSN. PIMRC 2014 - 25th Annual International Symposium on Personal, Indoor and Mobile Radio Communications, Sep 2014, Washington, United States. ⟨hal-01094472⟩