Reduce the Number of Sensors: Sensing Acoustic Emissions to Estimate Appliance Energy Usage

  • Authors:

  • Frank Englert;Irina Diaconita;Andreas Reinhardt;Alaa Alhamoud;Richard Meister;Lucas Backert;Ralf Steinmetz

  • Affiliations:

  • Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany;Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany;School of Computer Science and Engineering, University of New South Wales, Sydney, Australia;Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany;Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany;Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany;Multimedia Communications Lab, Technische Universität Darmstadt, Darmstadt, Germany

  • Venue:
  • Proceedings of the 5th ACM Workshop on Embedded Systems For Energy-Efficient Buildings
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

As a consequence of rising energy prices, manifold solutions to create user awareness for the unnecessary operation of electric appliances have emerged, e.g., real-time consumption displays or timer-based switchable wall outlets. A common attribute of these solutions is the need to buy and install additional hardware, although their acquisition costs often diminish the attainable savings. Furthermore these solutions only permit to retrieve accumulated figures of the energy consumption. Especially in households or office spaces with multiple persons, however, attributing electricity consumption to individuals provides enormous potential to determine possible savings. We therefore propose a system that allows to identify the energy demand incurred by a user's action based on audio recordings using smartphones. More precisely, we capture the user's ambient sounds and applying suitable filtering steps in order to determine the user's current activity. Our results indicate that our system is capable of detecting 16 typical household activities at an accuracy of 92%. By annotating the detectable household activities with information about typical energy consumptions, extracted from 950 real-world power consumption traces, a good estimate of the energy intensity of the users' lifestyles can be made. Our novel personalized energy monitoring system shows people their personal energy consumption, while maintaining their user comfort and relinquishing the need for additional hardware.