On the Black-Box Stand-by Recognition Strategies in Smart Homes Environments

The number of smart devices connected h24 to the electric plant of an house is continuously growing with a clear impact on the energy consumption footprint. In the near future the stand-by consumption of these appliances will represent about the 15% of the total house consumption, a remarkable percentage intended to increase due to the growing number of smart devices contained in the house of tomorrow. In the presented scenario smart houses play an important role to reduce the stand-by consumption footprint introducing new strategies able to autonomously recognize these energy wastes. In 1999 the International Energy Agency launched the one-watt initiative with the aim of fighting this problem, today's smart homes can contribute to reduce the phenomenon with a 60% saving. Different strategies can be used in order to recognize a stand-by state, they can be divided in supervised and unsupervised approaches. In this paper we present an algorithm for each kind of approach comparing the performance of the two single approaches with the aim of proving that unsupervised techniques are ready to be widely adopted. An experimental evaluation concludes the paper.