FRESH Research Platform

June 30th, 2014 § 0

Home energy management systems (HEMS) can reduce energy consumption while maintaining comfort. However, HEMS suffer from slow adoption and unproven energy savings. As intelligent automation and persuasive design features are increasingly added to HEM products, many research questions remain regarding their efficacy.

We present the FRESH research platform to facilitate
(i) the development of innovative features in HEM applications, and
(ii) their evaluation in field deployments.

FRESH has extensible modules for
(a) sensing to acquire data from wireless hardware sensor nodes and online sources, and actuation to control custom devices and commercial-off-the-shelf appliances,
(b) adaptive and intelligent modeling to test various machine learning and control algorithms,
(c) flexible user interface to explore the design of features based on behavioral research, and
(d) field experimentation with tools for user interaction logging, automatic context-aware surveys and messaging, and data visualization.

We have deployed FRESH in residential apartments and commercial offices used as test beds. We demonstrate the versatility of FRESH with four application use cases: environmental monitoring, non-intrusive load monitoring (NILM), an elevator energy display, and a smart thermostat. We have used FRESH to iterate on sensor nodes, develop and evaluate intelligent algorithms and models, and design mobile and web user interfaces to investigate HEM usability.

Publications & Presentations

  1. Lim, B. Y., Roth, K., Nambiar, S., Rayakota, H. 2014. Rapid Prototyping of Energy Management Applications with FRESH. In ACEEE Summer Study 2014.