Proposal for Energy Feedback Pilot Infrastructure Project Summer Banny Banerjee, Scott Klemmer, and Carrie Armel

Transcription

1 Background Proposal for Energy Feedback Pilot Infrastructure Project Summer 2009 Banny Banerjee, Scott Klemmer, and Carrie Armel Energy sensing technologies will be pervasive in the near future: electricity and gas smart meters are being installed en masse; home area networks (HANs) will be commercialized within the next year with plans underway to deploy them at scale; transportation sensors that quantify miles per gallon, mode (e.g., biking or driving), and number of trips (e.g., FastTrak) are available; and HVAC diagnostics derived from sensor data are being developed. In a different vein, work in public health and psychology shows that quantification serves as a powerful tool in promoting behavior change. This is achieved directly by providing specific feedback, such as on the amount of electricity one consumes 1, and also indirectly, by enabling incentive programs, markets, competitions, visualization, quantification in game and social networking applications, automated appliance control and behavior change guidance, and many other techniques. Coupled with a well-planned behavioral program, the billions being spent on sensor technologies can be leveraged to significantly reduce U.S. energy use; without such a program the investment is jeopardized because sensor information is complex and dull, and will be underutilized. The potential of this program may be derived from the potential of behavior change to reduce residential energy consumption which is estimated to be about 30 percent without waiting for new technologies, making major economic sacrifices, or losing a sense of well-being. 2 This is about 11 percent of the total U.S. energy consumption equivalent to the quantity of fossil fuels and CO2 emissions reduced by a 25-fold increase in wind plus solar power, or a doubling of nuclear power. 3 We propose such a behavioral program to leverage sensor data in order to achieve widespread energy reductions. We envision this program as having three parts: (I) a platform for prototyping, experimentation, and data collection, (II) an initial set of interventions and supporting work aimed at promoting energy reductions within individual households, with the work being founded upon powerful behavior change techniques, and (III) larger-scale analysis of behavior and HAN deployment through economic and engineering analyses. Through the platform, interventions and their components can be evaluated for their effectiveness quickly, easily, inexpensively, and at scale. This is possible because: (a) automatic generation and tracking of experimental manipulations is enabled because the interventions are implemented in electronic media such as an internet site and mobile applications, and (b) objective measures of behavior change are collected automatically by sensors that feed information into databases. The behavior change interventions will be implemented through the platform. At this point in time, we have made significant progress in developing a research plan for this behavioral program. This plan spans the work of numerous faculty and disciplines an overview is provided in the attached Figure 1. Development of the research plan formally began with a planning grant from PEEC and Woods Institute during the summer of 2008, and continued with workshops as well as meetings between faculty, students, and industry and government representatives. Several industry partners, such as Google and PG&E, are very interested in collaborating with us; this would enable us to test the platform and interventions with large numbers of subjects and additionally be well-

2 positioned to scale quickly. We have also identified and are planning to pursue several large grants to fund the effort, including those from NSF, ARPA-E, CPUC, The Energy Foundation, and perhaps others. However, these funds will take several months to be approved and dispersed. In the interim, several faculty and students are eager and willing to begin work in this area with a small amount of starter funds. Goals The current proposal requests funds for several of these researchers to achieve two goals: (1) Begin work setting up a platform for collection and display of sensor data, interface and intervention prototyping and evaluation, and experimentation (Part I described above). The end goal will be to have a working system composed of sensors in half a dozen homes, with a web application where individuals can view and change their energy usage. This will be achieved through the following implementation: We will install a dozen or so sensors/actuators per house in several houses. An embedded, selfcontained server will aggregate this information and send it to a web server. The sensors will talk to the server using something like XML/RSS. Javascript and PHP will be used to set up the web application. Jeff Heer's prefuse toolkit may be used for the graphics. (2) Begin developing and evaluating energy feedback interfaces (initiating the intervention development process described in Part II above). The goal is to gain intuitions about what constitutes an effective user experience for this domain, and to learn more about how changing the interface might change people's behavior. This will involve ethnographic research as well as prototyping, i.e., quick, iterative development and testing. Scope of Work The initial development of the infrastructure is to be carried out on the following fronts: 1. Monitoring hardware and data capture a. Evaluation of monitoring sensors and devices b. Selection and acquisition of sensors and monitors c. Calibrating sensors, establish sensitivity and linearity d. Installation of monitors and sensors in other floors of Donner dorm and 3 homes and in a lab setup e. Establish data link to server and population of database 2. Data processing, storage, and access a. Set up server, load software and set up on network b. Create prototypical database structure to accept data c. Establish data link to software d. Design software for preprocessing incoming data stream e. Create XML/RSS feeds and a data access layer 3. Developing and testing intial interface ideas (Brainstorms and rough prototypes) a. Develop crude prototypes on paper for multiple feedback scenarios in conjunction with specific behavior modification questions b. Develop multiple ideas for gaming and test them using crude prototypes

3 c. Create concepts on paper for computer interfaces as well as mobile platforms involving mobile applications, on board sensors, SMS, and Twitter 4. Platforms for feedback on the following platforms a. Prototypes of web based interface prototypes b. Prototype of interactive gaming prototype using Flash c. Rapid prototypes of top concepts from brainstorming phase 5. Automated control of hardware (this work will be carried out in Fall 2009) a. Acquire hardware for automated control of circuits, individual outlets and controllable thermostats b. Create prototype of software based control of outlets c. Create prototype of interactive feedback interface with control of outlets/thermostats built into the interface Work-Flow The group will meet once a week with the following timeline: Week 1: Get a house's worth of 2-3 alternative physical sensor systems, like PlugWise ( Widefield ( and Tendril. Have them plugged in by the end of the week. While waiting for them to arrive, explore existing webapps and make some "Hello World" programs to become familiarized with the constituent technologies, like PHP and RSS. Week 2: Make some architecture decisions. Which physical technology? Which home aggregator? What will you send to the cloud and how often? Use the HCI server as the cloud, or something else? Implement the web interface in PHP and Javascript? Week 3: Make paper, video, or other quick-and-dirty prototypes of the interface. Week 4: Have a house that can send data out over e.g., RSS Week 5: Store that data in the cloud. Week 6: Display that data on a web page in as quick-and-dirty a fashion as possible, e.g. all text. Weeks 7 & 8: Get some sort of graphical interface up and running. By the end of week 8, deploy it to 4-6 friendly users on their own data Weeks 9 & 10: Revise based on feedback from 4-6 friendly users. Post week 10: Ideally, write a CHI note/paper Personnel Banny Banerjee is Associate Professor in the Mechanical Engineering Department at Stanford and the Director of the Stanford Program in Design. He also heads the Stanford Design for Change Center. Prof Banerjee is involved in exploring transdisciplinary processes and paradigms that would cause large scale impact for global issues such as energy and climate change. He is interested in the ways in which "design thinking" can be used to create large-scale impact by combining innovation, technology, human behavior, new ecosystems for business, sustainability and the dynamics of rapid change. He has extensive experience in human centered design. He has a worked in the fields of architecture, design, structural engineering, energy, mechanical engineering, economic development work in emerging economies, sustainability, and software design. Scott Klemmer, Ph.D., is an Assistant Professor of Computer Science at Stanford University, where he co-directs the Human-Computer Interaction Group. Organizations

4 around the world use his lab s open-source design tools, and several books and popular press articles have covered his research. He believes that prototyping is pivotal to design innovation, collaboration, and creativity. His group s research empowers more users to design interactive systems, expert designers to be more creative, and programmers to engage in more design thinking. His research has introduced techniques for users to demonstrate interactive behavior, sample existing design elements to create new ones, and tightly integrate the creation and evaluation aspects of design. They explore these issues on mobile, desktop, web, and sensor-based platforms. They seek to understand the fundamental psychological and social ingredients of design excellence in order to create more effective design tools and practices. They create prototypes that envision future interface styles. Carrie Armel, Ph.D., is a Research Associate at Stanford s Precourt Energy Efficiency Center (PEEC) where she investigates how an understanding of human behavior leads to improvements in energy efficiency. Dr. Armel co-chairs the Behavior, Energy, and Climate Change Conference and oversees Precourt Institute s Behavior and Energy Bibliographic Database and Website. She has developed energy efficiency interventions that apply behavioral and design principles, and develops measures to evaluate the efficacy of such interventions. She holds a Ph.D. in Cognitive Neuroscience, which utilized behavioral, psychophysiological, and neuroscientific methods to investigate affect and motivation. She also performed post-doctoral work in Stanford s Economics Department and the School of Medicine. The team also includes several undergraduate and graduate students, including the PlugView team which was organized by undergraduate Kevin Mori. That group carried out a project in Spring quarter, directed by Banny Banerjee to build monitoring infrastructure for one floor of Donner Dorm. The team of six students installed metering hardware in the circuit box and demonstrated data capture at the level of each circuit with wireless communication to a database on the server, where the information was processed and made accessible as a feed to an interface dashboard such as Flash. This preliminary set-up was limited to circuit level data, one building, and a local server; the work in the current proposal would extend this work significantly in that it will receive data from plug-level sensors and send signals to control devices, receive data from many locations and make it accessible through the internet, ensure that the database it robust for larger data sets, and make the data accessible to prototyping tools beyond Flash. The Plugview team also proposed several preliminary interface concepts including modified switches and on screen interfaces which could enable gaming scenarios such as inter-dorm competitions. In the current proposal, this work will be significantly extended and complemented by ethnographic research to explore more ideas, develop initial prototypes, and receive feedback from users to refine interface concepts. In addition to the PlugView team, there are a number of students and recent graduates who are interested in working on infrastructure and design for energy feedback studies. References 1 Dena M. Bravata, Crystal Smith-Spangler, Vandana Sundaram, Allison L. Gienger, Nancy Lin, Robyn Lewis, Christopher D. Stave, Ingram Olkin, John R. Sirard. Using Pedometers to Increase Physical Activity and Improve Health: A Systematic Review. JAMA, 2007 Vol 298, No. 19

5 Darby, Sarah. (2006). 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. Residential Electricity Use Feedback: A Research Synthesis and Economic Framework. (2009). EPRI, Palo Alto, CA S. Nadel, A. Shipley, and R. N. Elliot, The Technical, Economic and Achievable Potential for Energy-Efficiency in the U.S. A Meta-Analysis of Recent Studies, Proceedings of the 2004 ACEEE Summer Study on Energy Efficiency in Buildings, Pacific Grove, CA, August 2004, (accessed 5 May 2008) D. Pimentel et al., US Energy Conservation and Efficiency: Benefits and Costs, Environment, Development and Sustainability 6, no. 3 (2004): U.S. Greenhouse Gas Abatement Mapping Initiative, Reducing U.S. Greenhouse Gas Emissions: How Much at What Cost? (Washington, DC: McKinsey & Company, 2007), /US_ghg_final_report.pdf (accessed 4 May 2008). American Physical Society (2008), Energy Future: Think Efficiency. Energy Future Report, pp Available at John A. Skip Laitner, Karen Ehrhardt-Martinez, & Vanessa McKinney. (2009). Examining the Scale of the Behaviour Energy Efficiency Continuum. American Council for an Energy-Efficient Economy, Paper ID Presented at the European Council for an Energy Efficient Economy Conference, 6/1/09, Cote d Azur, France. Gerald T. Gardner and Paul C. Stern. (2008). The Short List: The Most Effective Actions U.S. Households Can Take to Curb Climate Change. Environment Magazine, 50(5). 3 Sweeney, James (2007). Energy Efficiency Overview. Snowmass Workshop.