invited keynote on carbon, energy and the role of ubicomp tokyo-denki dec 2010

CARBON, ENERGY AND THE ROLE OF UBICOMP

Adrian FridayLancaster University

School of Computing and Communications

Thursday, 2 December 2010

OUTLINE

• Stimulating behaviour change w.r.t. sustainability and energy use is a hot topic

• Aim is to explore energy use and GhG externality further

• Present some of the well known exemplars from the Ubicomp literature

• Stimulate discussion on what might make a difference


FORTHCOMING ARTICLEIEEE Pervasive Special Issue on Smart Energy Systems


DEMAND MANAGEMENThttp://caniturniton.com/


http://caniturniton.com

http://caniturniton.com

CHANGING BEHAVIOUR


http://www.youtube.com/watch?v=gV7M9O1Wy3Y

http://www.youtube.com/watch?v=gV7M9O1Wy3Y

several concerns about the use of energy in his house. He

was particularly concerned with the amount of energy that

might be being used by household appliances left on

“standby” and explained that each night he would tend to

walk around the house and turn off all of the appliances

that he could that weren’t being used but had an LED light

showing that they were on stand-by. Ivan also worried

about whether boiling a kettle on the gas stove used more

or less energy than boiling an electric kettle.

3 Design

The aim of the domestic technology study was to identify

areas of domestic life that might be fruitful for ubiquitous

computing research. Having identified environmental

concerns and specifically concerns about energy con-

sumption as a possibly fruitful application area we

produced a series of possible designs for interfaces to

display energy consumption in the home, four of which are

shown in figure 1.

Fig. 1 Examples of initial design ideas for visualization of

energy consumption in the home.

We then went back to several of the households that had

been involved in our original study on experiences of

domestic technology and showed them a presentation of

these different ideas for representation of energy usage to

and asked for their feedback as well as more general ideas.

From these sessions, we decided to focus on the cuckoo

clock idea as a candidate interface for implementation and

began developing software to implement this interface.

One idea was that the cuckoo would ‘cuckoo’ every time 1

kilo of carbon was emitted to the atmosphere because of

electricity usage. Since cuckoo clocks are a German

invention (not Swiss as Orson Wells might have you

suppose), and the member of our team who was

implementing the interface was also German, we named

the project “Kuckuck” – the German word for Cuckoo.

4 Sensor Installation

To develop the prototype, we first set up a system to

monitor energy usage in the home via sensor-based

hardware and then trialled its usage in one of our

participant’s homes. We then developed an interim

representation to provoke discussions about how to

display the information. The emphasis at this stage is

not on behaviour change per se but on exploring the

sensors and displays that could be used in subsequent

studies to monitor bahaviour changes in energy

consumption. We describe our process to date in the

following.

4.1 Hardware Setup

For the hardware we decided to use a custom-built

electricity current sensor which can fit around the

live wire of the mains supply to a house. This is

connected to a particle Smart-It [3] which has been

loaded with custom software to increase the battery

life of the smart-it (from 12 hours to several weeks).

Custom-built water-temperature sensors were also

installed and attached to all of the pipes that flowed

in and out of the gas-fuelled hot water boiler. Again,

these sensors were attached to a particle Smart-It

which had custom software installed to ensure that

readings were only sent when there were significant

changes in temperature readings, thus radically

increasing the life of the particle Smart-It’s battery.

The particle Smart-It wirelessly broadcasts the sensor

readings and they are received by the particle Smart-

It data bridge which is attached to a laptop.

4.2 Real-World Installation Issues

Fig. 2 Service pipes in the immersion heater cabinet

where sensors were attached.

Having identified a family – Karen, Ivan and Josh -

who were willing to allow us to put electricity flow

and water temperature sensors in their home (figure

2), we embarked on a lengthy process of design and

test of sensor hardware, together with development of

health and safety, ethical and testing procedures. One

surprise for us was how much longer this process

took than we had originally anticipated, due to a

number of practical technical as well as logisitical

issues, pointing to the value of testing in authentic

environments. We have detailed this experience

elsewhere [2]. Having overcome many technical

obstacles, we have now managed to gather several

FEEDBACK DISPLAYSStringer, 2007


AMBIENT FEEDBACKGustafsson/ Martinez,2005

Broms, 2006

displays provide the feedback untimely or in a way that is difficult to understand. It requires the user to make mental efforts to translate the available information into appropriate actions. Furthermore, the information is not presented in the context where it is needed most i.e. when interacting with the home appliances or environment. Therefore the feedback lacks a direct and tangible link to the consumers’ behaviour. Current mechanisms also frequently have shortcomings with regard to long-term effectiveness, as initial results tend to wear off once the novelty effect is over [4].

PEEM therefore aims at improving the communication of energy feedback by seamlessly integrating it in the environment of the user and providing it where and when it is most useful and efficient. Such an integration of feedback could increase the comfort of the users, as no abstract translation and explicit attention towards achieving the goals is needed. Moreover, positive effects on the sustainability of behaviour change are expected.

The main starting point for the study is to explore persuasive technologies to influence behaviour towards optimized end-user energy management. Recent technological progress especially with regard to computational power, connectivity, availability of data and equipment cost allows deploying persuasive technologies in more and more contexts economically. Hence, advanced strategies of persuasion are technically possible. The potential of such approaches have been shown in different contexts (e.g. [2][3]). Within PEEM we aim to systematically explore the possibilities of ambient persuasive home displays for energy savings and develop targeted strategies for achieving energy-savings in the context home. PEEM

will build on knowledge from existing design concepts and approaches, three especially relevant examples are shown on the left.

The developed persuasive ambient displays will be experimentally evaluated in 30 households in and around the city of Salzburg. As we are interested in long-term effects the devices will be in use for 6 months in the households allowing us to differentiate between initial and sustainable effects.

The project will deliver valuable results on different levels. First, new prototypes and tools for providing situated and persuasive energy feedback will be developed. Second, guidelines on how to best implement ambient energy feedback in the home context will be defined and third, an empirical quantification of achievable effect sizes using persuasive ambient displays is determined.

References [1] Darby, S. 2006 The effectiveness of feedback on energy consumption. A review for DEFRA of the literature on metering, billing, and direct displays. University of Oxford, Environmental Change Institute.

[2] Froehlich, J., Dillahunt, T., Klasnja, P., Mankoff, J., Consolvo, S., Harrison, B., and Landay, J. A. 2009. UbiGreen: investigating a mobile tool for tracking and supporting green transportation habits. In Proc. CHI '09. ACM.

[3] Gustafsson, A. and Bång, M. 2008. Evaluation of a pervasive game for domestic energy engagement among teenagers. In Proc. ACE '08, ACM.

[4] Henryson, J., Håkansson, T. and Pyrko, J. 2000. Energy efficiency in buildings through information – Swedish perspective. Energy Policy. Volume 28, Issue 3

3 Examples for ambient energy feedback: Power Aware Cord (top), Energy Orb (middle), Energy AWARE Clock (bottom)

52












52












52


schedules, carpooling options, and data on the performance of their electric car.

MOBILE DEVICES The mobile application provides feedback and control to the residents of the home from their pocket – a simplified remote house control. The mobile application was developed for the Apple iPhone and iPod Touch (Figure 5) as an extension of the desktop web application and is currently being ported to other mobile platforms such as Symbian. It offers a subset of the features available in the web application, with each feature redesigned for use on a mobile device. For example, the controls available from the mobile emphasize ease of use through logical groupings. These “master” controls allow the resident to adjust the lights for a whole room, or shades for a whole house façade, with a single control. More fine-grained control of individual fixtures is still available, but a hierarchy of control makes the most commonly used items easily accessible.

INFORMATIVE ART ALIS also includes unique informative art approaches to providing feedback. The Ambient Canvas is one example of this approach: an informative art piece embedded in the kitchen backsplash (Figure 6). It gives feedback on the use of resources such as electricity, water, and natural gas. As opposed to typical graphical displays that may use numbers or charts to convey information, the Ambient Canvas combines LED lights and filters of various materials to produce light effects on the kitchen backsplash. This subtle feedback on performance and energy efficiency does not require active attention on the part of the resident, and integrates into the home cohesively. In

this context, informative art is intended to promote awareness of resource use to assist and influence sustainable in-the-moment decision-making.

Conclusion While we realize that information and ubiquitous technologies may play a powerful role in encouraging conservation, the design of these systems for effective home use faces critical challenges. We are exploring these design parameters through an integrated range of user interfaces on different platforms, from smart phones and web browsers to embedded and ambient displays, each designed to support sustainable decision-making in the home. We have operationally tested the system in two very challenging public showcases, where more than 100,000 people have visited and interacted with the live systems. From these experiences, we have already learned several important lessons:

1. Aesthetics are critical, both for those who build the environment (architects and interior designers) and those who live in them. We underestimated how

Figure 6: The “Ambient Canvas” kitchen backsplash display.

Figure 5: One screen of the ALIS Mobile application, showing a master light control for the Dining area.

23

REWARDING THE RIGHT BEHAVIOUR“The Ambient Canvas”, Bartram, 2010


shared goals; and a hierarchical model of energy-control settings to enable “one-step” optimization.

Controls As in standard home control, ALIS enables the resident to control and monitor lights, shades, and climate settings. In addition, the resident can configure energy-optimizing “modes” as presets in ALIS controls. For example, turning off most lights and lowering the thermostat in “Sleep mode”, or tuning settings and shutting down standby power in “Away” mode. These presets can be activated either by one button from any ALIS control interface (such as the mobile phone or embedded touch panel — Figure 2) or scheduled for planned activation. For example, in a prototype currently under development, a smart alarm clock by the bed can wake both the resident and the house (by putting the latter into Home mode). Note that these are presented as examples: modes are entirely user configurable, and coexist with individual control settings for fine-grained control when desired.

Controls differed slightly in North and West House(s). In North House, we added override controls for the sophisticated internal and external automated shade systems that tracked the sun, and extra state information to show the house was in “automated” or “manual” mode. Visitors to North House were intrigued by the efficiency of the automated shading system, but uncomfortable with the idea that if they wanted to change the behaviour (for example, to open shades to read a book) they had to suddenly “manage” the house control system. They struggled with a model of how the system worked, with what “optimal” and “non-optimal” modes represented, and with how they might balance their needs with the apparent state of the system. In a

different approach, West House currently uses no automated devices for climate control: we are experimenting with leaving energy optimization in the hands of its occupants, supported by contextual feedback to enable informed decision-making.

Feedback

Figure 3: The ALIS dashboard indicates daily usage statistics and provides uncomplicated data visualizations for at-a-glance awareness of resource consumption. It also conveys tips related to usage data and displays residents’ progress toward community challenge goals.

ALIS provides a variety of feedback displays and analytical tools. Detailed information on resource production and consumption is available in real-time and historical views, categorized in different ways (by type of device, by location in house, by time of use). We have integrated Pulse Energy™ software for detailed performance analysis and prediction (Figure 4). These detailed views complement an Overview

Figure 2: The embedded control panel interface from the garage in West House. The panel provides local control points for lighting and climate, and one-touch house mode buttons (upper right). Tabs to access other ALIS views are aligned along the bottom of the interface.

21

RAISING AWARENESSDEHEMS/Bartram, 2010


PRODUCTCurrentCost, DIY Kyoto, Enistic, e.g. http://www.diykyoto.com/


http://www.diykyoto.com

http://www.diykyoto.com

MAKING FEEDBACK WORK

• According to Midden, 1983

• Feedback should be immediate;

• concrete and significant (units, money); and,

• meaningful (one use or from comparable situation)

• McCalley, 2002 add

• salience (e.g. feedback integrated with a task, e.g. washing clothes)

• identify goal setting as highly effective (~20% savings)


Continuous noise is usually intrinsic to the device's operation and internal electronics. Appliances like grinders, fans and hair dryers that make use of a motor create voltage noise synchronous to the frequency of AC power (60 Hz in the USA) and its harmonics (120Hz, 180Hz, etc.) due to the continuous making and breaking of electrical contact by the motor bushes. In contrast, modern SMPS based electronic devices generate noise that is synchronous to their power supply’s internal oscillator.

In contrast to traditional linear power regulation, a SMPS does not dissipate excess power as heat, but instead stores energy in an inductance and switches this stored energy in from the line and out to the load as required, thus wasting much less energy. The key to a SMPS’s smaller size and efficiency is its use of a power transistor to switch the stored energy at a high frequency, also known as the switching frequency. The switching frequency is much higher than the 60Hz AC line frequency because at higher frequencies the inductors or transformers required are much smaller [7]. Typical SMPS operate at tens to hundreds of kHz. The switching waveform is adjusted to match the power requirements of the appliance it is powering.

Figure 1: (Left) Circuit model of a SMPS with placement of

the voltage probe. (Right) Frequency domain analysis at the

voltage from probe showing EMI at 10 KHz.

A CFLs power supply employs the same fundamental switching mechanism to generate high voltages necessary to power the lamp. The switching action, which is the cornerstone of a SMPS’s operating principle, generates a large amount of EMI centered in frequency around the switching frequency. This phenomenon can be understood by modeling a simple DC-DC SMPS circuit that uses the same fundamental switching topology. See Figure 1 for schematic.

The large inductor L_PowerLine mimics the power line inductance. The SMPS is plugged into the power line. To measure the conducted EMI, we place a voltage probe V on the power line, which is analogous to having the single sensor plugged into the power line with a SMPS based device operational somewhere else. The switching frequency fc for the model is governed by the PER (period) parameter of the V_Switching component. We arbitrarily set it to 10 KHz. Figure 1 shows a frequency domain plot of the noise at probe, which clearly shows that the power supply emits EMI, which is conducted over to the power line and is most prominent at the switching frequency fc (10 kHz here) and its harmonics. This is the same behavior that

we observe when a SMPS based appliance is turned on in the home.

In the US, the Federal Communications Commission (FCC) sets rules (47CFR part 15/18 Consumer Emission Limits) for any device that connects to the power line, which dictates the maximum amount of EMI a device can conduct back onto the power line. This limit is 66 dBuV for frequency range between 150 kHz to 500 kHz, which is nearly -40dBm across a 50 ohm load. The ElectriSense data acquisition system is sensitive enough to capture noise from -100 dBm to -10 dBm across a frequency range of 36KHz - 500KHz.

Figure 2: Frequency spectrogram showing device actuation in

a home.

Figure 2 shows a frequency domain waterfall plot showing appliances being turned on and off. As is evident from the graph, when the device is turned on we see a narrowband continuous noise signature that lasts for the duration of the device’s operation. Also note that the noise center is strongest in intensity and then extends to lower and higher frequencies with decaying intensity, which can loosely be modeled with a Gaussian function having its mean at the switching frequency. This behavior can be attributed to the error tolerance of the components that make up the switching circuit core, as well as the characteristics of the power supply's load. If all the components were ideal, we would see a single narrow signal peak at the switching frequency. The error tolerance of SMPS components also allows for distinction between otherwise identical devices, such as a variety of units of the same model of CFL bulb. Finally, the power line itself can be thought of as a transfer function (difference in the inductance between the sensing source and the appliance) and provide additional discrimination among multiple similar devices. We show this experimentally later in this paper.

Dimmers also produce continuous noise due to the triggering of their internal triac switches, which can be used to detect and identify incandescent loads they control. In contrast to the narrowband noise produced by SMPS, a dimmer produces broadband noise spanning hundreds of kHz, which could be modeled as a Gaussian having very large variance. A more detailed treatment of dimmers and distinction between identical devices is presented also presented later in this paper.

DISAGGREGATION USING SINGLE POINT SENSINGElectrisense, Patel 2010


household once smart meters are introduced. We wanted the user interface to be attractive, easily-accessible and to provide functionality that motivates the user for a long period. Thus, our application pro-vides the following major functionalities. It shows the overall energy consumption of the household in real time (fig. 1 left) as well as the historical consumption on a daily, monthly, or weekly basis. Furthermore, the system allows for the user to interactively measure the consumption of any appliance or set of appliances in the house and thereafter allows personalization of the measurement. Finally, it shows an appliance summary that provides an overview of the consumption, the costs, and corresponding equivalents per measured appliance (fig. 1 right).

The backend architecture of our prototype is based on three independent elements (fig. 2). The first monitors and logs the energy consumption by the sensors of the smart meter. The second element, the gateway, con-sists of a parser, a database, and a tiny web server. To acquire the logged data from the Landis + Gyr smart meter E750 on a continuous basis in real time, the Smart Message Language (SML) [3] parser automati-cally polls the meter’s

figure 1. User interface accessing real-time metering data

data and stores it in a SQL database. In order to enable interoperability with other applications, the web server offers access to the gateway’s functionality and the smart meter’s sensor values using URLs. This approach originates from the Web of Things [4], where connec-tivity to the functionalities of real-world devices is ex-posed using a REST API. The last element is the visuali-zation interface. It uses the functionality provided by the gateway to access the data base and to dynamically present real-time information about the energy con-sumption.

Future work will address in more detail an evaluation of the prototype and an analysis of the potential to auto-matically recognize previously measured appliances.

figure 2. Smart meter communicating with the mobile UI

References [1] Petersen, J., et al. Dormitory residents reduce elec-tricity consumption when exposed to real-time visual feedback and incentives. Int. J. of Sustainability in Higher Education (2007), 16-33.

[2] Abrahamse, W., et al. A review of intervention studies aimed at household energy conservation. J. of Environmental Psychology, 25 (2005), 273-291.

[3] Smart Message Language. http://www.t-l-z.org/docs/SML_080711_102_eng.pdf.

[4] Guinard, D., et al. Towards Physical Mashups in the Web of Things. Proc. INSS, IEEE Press (2009). MOBILE FEEDBACK

Weiss, 2009


SUSTAINABLE BEHAVIOUR?


http://www.youtube.com/watch?v=3iqI786nOEM

http://www.youtube.com/watch?v=3iqI786nOEM

CARBON VS. ENERGYGhG externality is a time varying phenomenon


REAL TIME CARBONhttp://realtimecarbon.org/


http://realtimecarbon.org

http://realtimecarbon.org

Fig. 2 The prototype Pollution e-Sign

Fig. 3 A m

an waiting near our sign inspects the B

luetooth message he has received ...

M

oments later a w

oman receives another m

essage which they com

pare

Fig. 4 The bluejacking device logs successful and

unsuccessful transmissions

POLLUTION E-SIGN

• What role for Ubicomp then?

• To measure, inform, nag, share, embarass, challenge, engage in play, stimulate enquiry?

• Hooker, 2007


MAKING AN IMPACT

• Typical UK person emits 15 tonnes CO2e

Household fuel 13%

Household Vehicle fuel

10%

Household electricity

9%

Personal air travel

8%

Other Personal transport

3%

Cars 5%

Food and drink (from

shops) 12% Hotels, pubs

and catering 4%

Paper and printing

1%

Textiles and clothes

2%

Electronic / computers / appliances

4%

Construction 6%

Water and Sewage

2%

Defence, education and

health and social services

11%

Other 10%


HONG KONG RETURN:4.6 TONNES

!"#$$%&'$("))*+$ ,-./$0%&'$("))*+$12*34.*$$'&5$("))*+$

67")"89$ :-3+($7;4++$


ASPARAGUS (250G PACK):2KG

!"#$%&'$($

)*+$,+-*(./-0$,+"1$2-+3$

)4-+5(-$&$6($

!"758$9:;<-5<":$

=*(.$>?'$6($


A RED ROSE:350G

!"#$%

&$'()$*+",%-.%&$//0.,%

12345%6/$7.%8#$9%:".;0%

6#$9%;$*#%40#,".%

&-4)%<=>%?4%


A BURGER:2.5KG

!"#$%&'$

()'*$+,-&'$

./012'0$3,-$&'$

40'')0$561'01$

7$89$:*00;0$<61'01$


IDENTIFYING TARGETS

• Reduce flights - better communication?

• Change your commute/ cut your mileage/ avoid congestion

• Get an efficient small car next time or ride share

• Cut food waste, seasonal food, food with low carbon miles

• Go vegetarian, eat less meat

• Simple efficiency measures (insulate, cut drafts, boiler, lightbulbs) - use energy more effectively, cut waste

• Buy less, buy quality, by more locally, look after it, mend it, freecycle it


MORE EFFICIENT BUSINESSES?

• Online sales account for 10% of the total retail sales in the UK

• 65 million online purchases (12%) weren’t delivered first time, with 2% failing to be delivered at all

• £682 million of direct costs will be borne by consumers, retailers and carriers due to Internet shopping delivery inefficiencies (£1.26 per purchase)

• Can we exploit delivery to one’s social network?

• Can Ubicomp help us be more aware of the downstream impact of our choices and behaviours?


THANKS TOMike Berners-Lee, Small World Consulting

http://bit.ly/9gBwDt




invited keynote on carbon, energy and the role of ubicomp tokyo-denki dec 2010

Documents

energy feedback

feedback displays of

com of energy feedback

energy use

variety of feedback

feedback work

energy consumption

ambientno abenergy feedback