Demonstration Of A Monitoring Lamp To Visualize The Energy Consumption In Houses

Abstract

We report on the development of a wireless lamp dedicated to the feedback of energy consumption. The principle is to provide a simple and intuitive feedback to residents through color variations of the lamp depending on the amount of energy consumed in a house. Our system is demonstrated on the basis of inexpensive components piloted by a gateway storing and processing the energy data in a WoT framework. Different versions of the color choosing algorithm are also presented.

Full text

Demonstration Of A Monitoring Lamp To
Visualize The Energy Consumption In Houses
Christophe Gisler1,2, Grazia Barchi1,3, G´erˆome Bovet1,
Elena Mugellini1,2, and Jean Hennebert1,2
1ICT Institute, HES-SO//Fribourg, Bd. de P´erolles 80, CH-1705 Fribourg
2Dpt of Informatics, University of Fribourg, Bd. de P´erolles 90, CH-1700 Fribourg
3Dpt of Electronic and Information Eng., University of Perugia
Via G. Duranti 93, I-06125 Perugia, Italy
christophe.gisler@hefr.ch
http://www.eia-fr.ch,http://www.unifr.ch,http://www.unipg.it
Abstract. We report on the development of a wireless lamp dedicated to
the feedback of energy consumption. The principle is to provide a simple
and intuitive feedback to residents through color variations of the lamp
depending on the amount of energy consumed in a house. Our system
is demonstrated on the basis of inexpensive components piloted by a
gateway storing and processing the energy data in a WoT framework.
Different versions of the color choosing algorithm are also presented.
Keywords: Web of Things, energy feedback, Green Computing
1 Introduction and motivations
Citizens are more and more sensitive to the environmental dimension linked to
their behaviours. Beyond the current ecological trends, interests are also eco-
nomical due to the raising prices of energy. It is reported in the literature that
energy feedback is a good solution to make people gain awareness of their energy
consumption and hence reduce it [1]. An efficient feedback should be grounded
in real consumption, provided with a short delay and include opportunities for
historical and/or social comparison [2]. Unfortunately, nowadays the unique feed-
back that most of the people have is their monthly electricity bill, which provides
few information to interpret the sources of consumption. Automated monitoring
of the electricity consumption in a house is quite a recent topic. The planned
introduction of smart meters in our homes and the availability of plug-based en-
ergy monitoring devices are pushing in this direction. Recent publications in the
domain report on several mobile and web-based energy monitoring systems [3,
4]. The major disadvantages of such kinds of systems are a strong dependence
to web connections and an active user implication in order to use them.
We describe here a simple and intuitive system able to give feedbacks on
energy consumption in habitations. The system is based on a wireless Energy
Monitory Lamp that provides visual feedback to residents through variations of
its color. The chosen color depends on the real-time amount of energy consumed
The 10th International Conference on Pervasive Computing (Pervasive2012).

2 Energy Monitoring Lamp
in the house allowing the residents to immediately adapt their use of appliances
in order to save energy and money. There are various kinds of feedbacks. Our
approach belongs to the class of ambient feedbacks which rely on pre-attentive
processing of information (i.e. the human ability to process rich information
without cognitive effort). Such devices do not show text or numbers, but simply
alert residents that something relevant to their electricity consumption is chang-
ing. Our proposal is actually very similar to the commercial device Orb which
changes colour according to the current price of electricity [5]. We differentiate
ourselves by documenting different versions of the algorithms to chose the color
of the lamp and by using open and inexpensive electronic components piloted by
a gateway storing and processing the energy data in a Web of Things framework.
2 System description
2.1 Architecture
Our system is actually composed of three communicating parts (see Figure 1):
an RGB LED-based lamp, a small-footprint gateway for the acquisition and
processing of the data and smart outlets measuring the electricity consumed by
various appliances in a building.
SQL
Database
Smart Outlets or
Power Meter
RGB LED-based Lamp
with Arduino µC
Energy Hub (PC)
ZigBee Bluetooth
Fig. 1. Schema of the system
RGB LED-based lamp. The lamp is controlled by an Arduino micro-
controller which is an open-source electronics prototyping platform. Such a card
can receive inputs from various sensors and is usually used to control lights, mo-
tors and actuators. It can also store information into a small EEPROM mem-
ory. The Arduino technology simplifies the process of working with a micro-
controller and presents some advantages, as it is cheap, multi-platform, exten-
sible, in an open software and hardware paradigm. Its hardware is based on
Atmel’s A TMEGA8 and A TMEGA168 micro-controllers. Among the available
model cards, we chose the ArduinoBT which has an integrated Bluetooth mod-
ule. The second electronic part of the lamp is a set of four RBG LEDs. For the
The 10th International Conference on Pervasive Computing (Pervasive2012).

Energy Monitoring Lamp 3
micro-controller to handle the RGB LEDs and make them show suitable color
transitions and intensity variations, we used a Pulse Width Modulation (PWM)
technique.
Smart gateway. The lamp communicates via Bluetooth with a small foot-
print PC called Energy Hub. This energy hub acts as a smart gateway between
the lamp and the sensors by getting the consumption data from the sensors,
storing them into a MySQL database. The hub is implemented using an open-
source platform called Watt-ICT that offers a WoT access to different brands of
smart outlets and building sensors/actuators (http://wattict.com/).
Smart outlets. As sensors, we used PLOGGs smart outlets that transmit
the measured electric consumption values to the energy hub via the Zigbee wire-
less protocol (http://www.plogg.co.uk).
2.2 Algorithms
Our objective here is to find a mathematical function for mapping the electricity
consumption at an instant tto a convenient color between green and red which
intuitively represents a low and high consumption. Consumption levels are con-
text dependent, i.e. what is high for a household may be low for another one.
For this reason, two approaches can be followed. The first approach, that we
named intrinsic, only relies on internal measurements coming from the installed
system. The second one, that we named extrinsic, uses external data such as
the number of residents, the type of building, the current weather or national
statistics about consumption. The best way to proceed is probably the combi-
nation of both approaches. In this paper and as initial work, we chose to focus
on intrinsic approaches in order to have a functional autonomous system. Two
algorithms have been implemented.
The first algorithm is based on the momentary power consumption P(t) com-
puted as the average of the instantaneous consumption over a sliding window.
The length ∆t of the window is user settable with a reasonable default equals to
30 seconds. The momentary power P(t) is then compared to the min and max
values that are set either automatically from measured extrema or manually by
the user. The output color for P(t) is chosen relatively to these extrema. Figure 2
illustrates the concept.
The second algorithm is based on the accumulated daily energy consumed
E(t) until time tand compared to the previous average ¯
E(t) over the last ndays.
The difference ∆E(t) = E(t)−¯
E(t) is then computed and if ∆E(t) = 0, we can
assume that the energy consumed is stable, leading to a neutral lamp color, i.e.
yellow as the mean between green and red in the HSV scale. We then compare
∆E to Erem(t) = ¯
Etot −E(t), the remaining energy quota to reach the total
daily average energy ¯
Etot. If ∆E ≥Erem (t), then the energy consumed until the
current time thas already exceeded what was observed during the previous days
and the lamp color is red. If ∆E ≤ −Erem (t), then we can expect that we will
reach a lower consumption in comparison to the other days, leading to a green
color of the lamp. Figure 3 illustrates this concept.
The 10th International Conference on Pervasive Computing (Pervasive2012).

4 Energy Monitoring Lamp
t = now
t-∆t Time
Power
Output
Sup.
Inf.
Mean
∆t
Current daily
consumption
24:000:00
Fig. 2. Feedback algorithm 1 - Ouput color computed from current momentary con-
sumption
Time
Power
t = now
Output
> Erem
< -Erem
∆E
∆E =
0
Current & mean
daily consumptions
24:000:00
Fig. 3. Algorithm 2 based on the difference between current and mean consumptions
3 Conclusion
The purpose of the energy monitoring lamp presented in this paper is to pro-
vide users with an intuitive and ambient feedback about their current energy
consumption. We detailed the three components of the system, namely an RGB
LED-based lamp piloted by an Arduino microcontroller, an energy hub process-
ing data and smart sensors measuring the electricity consumption. We presented
two algorithms used to model the household energy consumption and make the
lamp color vary from green to red accordingly.
References
1. S. Darby, The Effectiveness Of Feedback On Energy Consumption - A review for
DEFRA of the literature on metering, billing, and direct displays, Environmental
Change Institute, University Of Oxford, April 2006
2. C. Fischer, Feedback on household electricity consumption: a tool for saving energy?,
Energy Efficiency, Springer Science + Business Media B.V., 2008
3. M. Weiss, eMeter : An interactive energy monitor, Higher Education, 2009, pp. 3-4
4. D. Guinard, M. Weiss, and V. Trifa, Are you Energy-Efficient : Sense it on the
Web!, International Journal, pp. 3-6
5. M. S. Martinez and C. R. Geltz, Utilizing a pre-attentive technology for modifying
customer energy usage, Proc. EU Council for an Energy-Efficient Economy, 2005
The 10th International Conference on Pervasive Computing (Pervasive2012).