Toward Visualising and Controlling Household Electrical Appliances

Latha Karthigaa Murugesan Department of Electrical and

Computer Engineering The University of Auckland,

New Zealand lmur778@aucklanduni.ac.nz

Rashina Hoda Department of Electrical and

Computer Engineering The University of Auckland,

New Zealand r.hoda@auckland.ac.nz

Zoran Salcic Department of Electrical and

Computer Engineering The University of Auckland,

New Zealand z.salcic@auckland.ac.nz

Abstract—Energy consumption is a vital component in day-to-day life. Visualising energy consumption, recommending energy conserving strategies and, controlling the appliances are some of the widely considered means to motivate and/or to help end-users at household to conserve energy. However, the conventional energy visualisation applications have certain drawbacks, such as lack of energy consumption location, lack of controlling appliances based on energy-saving recommendations, etc. To overcome those, this research would accomplish a simulation model by designing the following: (i) visualisation engine, to visualise household electricity consumption, (ii) recommendation engine, to provide energy-saving recommendations to the end-users, (iii) control engine, to control the household appliances based on the recommendations, and (iv) integrate visualisation, recommendation and control engines. This research uses a combination of user-centred design and iterative development methods. Successful implementation of these objectives would provide a significant contribution to this inter-disciplinary research area and would serve to achieve higher energy savings.

Keywords-visualisation, power, household, energy consumption, system

I. INTRODUCTION Visualisation of energy consumption is a pictorial/animated representation of information about household energy use, accessible in digital formats such as PCs, laptops, smart phones, tablets, etc. [1, 2]. It provides information such as near-real-time power data (e.g., total power consumed by all the household appliances) [3, 4] archived energy data (e.g., energy consumed during the last week, month, etc.) [5-7]. The major advantages of visualisation are that the feedback on energy consumption is immediate and it is easy for the end-users to make energy saving decisions [8-10]. The visualisation incorporated with the recommendations on energy saving also helps the end-users to respond immediately to the energy behaviour [9, 11, 12]. Though the visualisation works well with recommendation framework, the system would be considered a failure when the end-users are unable to control the appliances based on recommendations (For example, the end-user is in office and receives notification/recommendation from the system that the air conditioner is switched on unnecessarily and that they may wish to turn it off, but the end-user is unable to control the appliance from their office). Hence, the aim of the research is

to develop a novel integrated visualisation and control system (VCEC) which visualises, recommends and allows users to control the household appliances to maximise energy savings by reducing consumption.

II. RELATED WORKS The related works targeted both the research outputs and the commercial solutions to identify the gaps in the literature and markets. The results are described in the following subsection.

A. Research Literature The literature survey results are categorised based on the

three functional requirements of the system. They are (a) visualisation of electricity consumption, (b) energy saving recommendations, and (c) controlling household appliances.

1) Visualisation of electricity consumption This subsection explains the information displayed in the

visualisation of energy consumption and it provides near-real-time power information for their immediate action, thus resulting in effective energy conservation [13-15]. In addition, peak energy consumption in a day or month or year has also been displayed [12].

Some studies [16, 17] compare and visualise the energy consumption during the homologous periods (i.e., hours, days, months, years, etc.) and the information has been displayed in the form of graphs or charts. This kind of visualisation helps end-users in identifying peak energy consumption periods. In addition to this, end-users were motivated by displaying their energy consumption in comparison to that of their neighbourhood buildings [16, 18].

Dominguez et. al. [16] researched the prediction of energy consumption (i.e., prediction of monthly bills) and verification of energy bills (i.e., validating it with the historical energy data).

2) Energy Saving Recommendation This section reviews the research literature describing the

various recommendations provided by the energy monitoring software for energy saving [6]. The recommendations are the smart energy tips or notifications provided by the software to impart energy conservation behaviour of the end-users. There were several papers targeting this aspect as the recommendation reaches end-users immediately and the decision making would be immediate to conserve energy.

There are 2 types of recommendations found in the research literature. They are (a) behaviour-based [19-21], and (b) environment-based.

The behaviour-based recommendation could be defined as the smart energy tips provided to the end-users based on the behaviour of household energy consumption. For instance, the studies [19, 20] explained the importance of providing the energy tips on device-level energy consumption, e.g., ‘The TV consumed more electricity than the previous week, switch it off when not in use’. The study [20] also discusses several templates for the energy saving recommendations based on energy consumption behaviour. The recommendations were also been made with cumulative energy usage [20] (e.g., The device ‘x’ has been turned on for 84 hours continuously in the past week) and, comparison with past energy consumption [19] (e.g., consumption during the current week is higher than the previous week by 5%).

The environment-based recommendation are the energy tips offered to the end-users based on environmental conditions such as temperature, pressure, humidity, light intensity, etc. Other studies such as [7, 20] investigated and visualised the energy consumption based on environmental conditions (e.g., using sensors). For example, suggestion was made to switch off the light when the light intensity was sensed to be too high.

3) Controlling household appliances This section surveys the research literature to present

information on controlling the household appliances. There are numerous research patents [22] and literature [23] explaining the importance of intelligent control of household appliances. Wong, et.al. introduced a phone-based system for home and office automation using a hardware-based remote controller for home appliances, where the communication takes place via a dedicated telephone line. A similar system has been designed for remote home automation and control using the telephone by Coskun, et.al. [24]. Few studies proposed automation system that can control home appliances from a PC using Bluetooth. One of the studies [25] explained the Java based monitoring and control systems which could control the appliances through the internet. The studies presented in [26, 27] have enriched the field of web-based real time applications and the presented systems are able to control, monitor, and interact with real devices.

B. Commercial Solutions There are very few commercial applications available in the market for visualising household energy consumption. This section would explain briefly about the literature in the commercial applications.

Visualisation of electricity consumption - The various information displayed in commercial applications are real time electricity consumption [3, 28], energy prediction, goal setting [28], historical energy consumption, disaggregation (appliance level energy information), displaying energy waste, peak energy consumption, voice enabled service [3], etc.

Energy saving recommendation - Commercial solutions also supports both behaviour [3, 28] and environment-based recommendations [3].

Controlling household appliances - Several small-scale applications had been developed to control the household appliances using internet through the web. Slowly, this moved to smartphone application in the last decade. Currently, there are a large number of Android and iOS applications available in the market such as Control your home, Home control assistant, iNELS home control mobile, etc. to control the appliances online.

C. Research Gap From the research literature and commercial solutions, the

following research gaps have been identified. • Current research on visualisation is limited to providing

aggregated & disaggregated energy consumption, near-real-time power data, energy prediction, etc., but lacks presenting end-users the information about where exactly at home (e.g., kitchen, living room, etc.) the energy is being consumed.

• There is a lack of research focus on recommendation for energy conservation. Existing literature provides recommendations based on either usage patterns (consumption behaviour) or context-aware data (using various sensors). However, no research has focused in detail on developing a recommendation engine based on both usage patterns and context-aware data and does not take into account the user preferences.

• Current research on control does not include controlling the appliances based on the recommendations and the user preferences such as timing constraints, targeted energy consumption, etc.

• There exists no single system that integrates these 3 aspects (visualisation, recommendation, control) across these 3 different research areas (human computer interaction, software engineering, and electrical engineering) for energy conservation.

III. RESEARCH OBJECTIVES The primary objective of the current research is to design

and develop the VCEC simulation model (henceforth referred to as the model). In the future, we hope to extend the model to a real system that works through real appliances and sensors. The research objectives (RO) to accomplish the VCEC model are explained below. RO1: To create usable and useful visualisation such that the end-users are better informed about the whereabouts of household energy consumption. RO2: To create a novel context-aware recommendation engine such that the system could make recommendations to the end-users based on usage patterns, context-aware data and user preferences. RO3: To create a unique intelligent control engine such that the end-users could control the household appliances based on

recommendations, timing constraints and targeted energy consumption. RO4: To create a novel, comprehensive VCEC model by integrating visualisation, recommendation and control engines.

IV. RESEARCH AND DEVELOPMENT METHODOLOGY The three subsystems of the VCEC system are visualisation, recommendation and control engines. This research uses the combination of Agile methods for software development, and UCD (User-Centred Design)/PD (Participatory Design) for design. The research and development methodology is summarized in Figure 1. The list of steps that is needed to develop a subsystem includes subsystem investigation, planning, design, implementation and evaluation [29].

Figure 1: Customised Agile Software Development Process

A. Subsystem Investigation In this step, the end-users and domain experts are to be

interviewed to understand the requirements of system. The data from the interviews would be analysed. The synthesised data yields a list of all features in the system.

B. Planning The planning step includes planning for the participatory

design workshop, developing sprints and its tasks, writing test cases and plan for usability evaluation resulting in an updated backlog, design plan and test plan document. C. Design

The planning is followed by the design with 2 sub-tasks, (a) participatory design, and (b) low fidelity prototyping. • Participatory design (PD): This step requires the

involvement of all the participants such as end-users, domain experts, etc. in the design process [30]. In PD, the users perform a number of activities, so as to make sure that an effectively designed system would be developed.

• Low fidelity prototyping: Low fidelity prototype is the mock-up or replica of the final developed product to check the correctness of the design.

D. Implementation and Evaluation The implementation involves developing the software code to specify the functionality and testing investigates the correctness of the software code and the software functionalities.

V. SYSTEM ARCHITECTURE AND SOFTWARE DESIGN The target device for visualisation is the 7 to 10 inch tablet

and the target platform or OS is Android. This section explains the overall architecture for visualising the VCEC system and is summarised in the Figure 2. This architecture is a client-server architecture, in which the server side comprises 4 layers: (a) input layer, (b) data layer, (c) application layer, and (d) presentation layer. The information about each layer is explained below.

Figure 2: Overall Software System Architecture

Input layer: The input layer is comprised of (a) sensors, to read the quasi-real-time (i.e., near-real-time) aggregated power data, (b) appliance data, which includes the information such as number and type of appliances, its energy rating, etc., (c) context data, which is the data to observe the environmental conditions such as temperature, light intensity, etc. using the respective sensors, and (d) user preferences, which includes targeted energy consumption for the particular week or month, timing constraint to convey whether a particular appliance could be operated at a particular time, and recommendation settings to choose either automatic or manual mode. Data layer: The data layer is comprised of (a) quasi-real-time power database (DB) to store the quasi-real-time power data and the database gets updated every minute, (b) appliance DB, which has the information about all the household appliances and its respective expected-power-consumption (in watts or kilo-watts), and (c) sensor DB, which helps store the information from miscellaneous sensors such as temperature, light intensity. Application layer: The application layer, which receives input from the database layer, comprises of three subsystems: (a) visualisation, (b) recommendation engine, and (c) control engine which are explained in the next subsection. Presentation layer: The presentation layer helps in creating the web pages or application screens suitable for end-users to view on PC, laptop, smart phones, etc. The information to be visualised are: (a) aggregated quasi-real-time energy data, (b) archived aggregated energy data, (c) weekly/ monthly energy

forecast (i.e., predicting the future weekly/ monthly energy consumption), (d) short and long term recommendations for energy conservation, and (e) suggestion to control household appliances.

A. Visualisation Engine This engine represents the information provided to the end-users for visualising household energy consumption. The visualisation in Figure 3 shows all the appliances and their corresponding energy information with the floor plan of the home to know the whereabouts of energy consumption. The information provided in visualisation are (a) aggregated and disaggregated quasi-real-time power data, (b) archived aggregated and disaggregated energy data, and (c) weekly/ monthly energy forecast. The label A in the figure represents the quasi-real-time power data, label B represents the archived energy data and label C represents the energy forecast.

Figure 3: Visualisation Engine depicted in Low Fi Prototype Drawing

• Aggregated and disaggregated quasi-real-time energy data

(Figure 3, Label A) • Archived aggregated and disaggregated energy data

(Figure 3, Label B) • Weekly/monthly energy forecast (Figure 3, Label C)

B. Recommendation Engine The recommendation engine could make recommendations

to the end-users to ensure energy conservation. A number of recommendation templates would be created based on which the recommendations would be mapped and presented to the end-users, e.g. template: ‘TV consumed $x this week’. The recommendation is provided to the end-users based on usage patterns, context-aware data and user preferences. The usage patterns represent the energy consumption patterns at household, e.g., the recommendation, ‘TV consumed more energy than the last month’ is based on the usage patterns. The context-aware data, in this context, is the information from the sensors such as temperature and light intensity sensors, installed at home and the recommendation would be generated based on the sensor data, e.g. the recommendation ‘switch off

the lamp as the light intensity is high’ is based on context-aware data. With the user preferences, the end-users could prefer/not prefer certain templates to appear in the visualisation.

In Figure 2, based on the data from appliance DB, sensor DB and the information from the visualisation, certain energy saving recommendations could be conveyed to the end-user either in short or long term. For example, the short-term energy recommendation includes informing the end-user about the unnecessary use of any appliance and then recommending them to switch it off. The long-term energy recommendation could be ‘by avoiding the usage of electrical appliances at peak hours would save an average of $x per year’.

Figure 4: Recommendation and Control Engine depicted in Low Fi

Prototype Drawing

C. Control Engine The main objective is to create a unique intelligent control engine such that the end-users could control the household appliances. The inputs to the control engine are the output of the recommendation engine, sensor data and the user preferences. There are three user preferences for controlling the appliances: (a) targeted energy consumption for the particular week or month (e.g. the end-user could set a target value for weekly/monthly energy consumption), (b) timing constraint to convey whether a particular appliance could be operated at a particular time (e.g. some people may not like operating the washing machine during night time), and (c) recommendation settings to choose either automatic or manual (e.g., the recommendation does not wait for end-users' permission in automatic mode, while manual mode does need permission to continue). This would result in reducing the energy consumption. Figure 4 represents the low fi prototype drawing of the recommendation provided to the end-users and also helping the end-users in controlling the appliances.

VI. INTIAL IMPLEMENTATION As previously cited, the VCEC system has 3 planned engines: visualisation, recommendation and control engines. The initial step in the implementation concentrated on the visualisation engine, where the interactive 2D visualisation of archived

aggregated household energy consumption has been created. The reason for developing this intermediate step (2D interactive visualisation) is to serve as a gap between the traditional paper bill and final interactive visualisation through easy accessibility of archived energy consumption. This implementation involved 5 different categories. They are (a) power consumption of an hour (every minute) – Figure. 5a, (b) energy consumption of a day (every hour) – Figure. 5b, (c) energy consumption of a month (every day) – Figure. 5c, (d) energy consumption of a year (every month) – Figure. 5d, and (e) comparison of energy consumption of two years – Figure. 5e. For proof of concept, the current implementation uses the aggregated energy data of a household in Finland from December 2006 to February 2008. In this application, the users can select any of these 5 categories from a drop down menu and provide details such as date and year to view the 2D interactive visualisation.

Figure 5. 2D Interactive Visualisation of Finland Household Energy Data

Almost 90% of NZ homes receive the electricity bill [31] that has 3 main information: kWh, cost (in NZD), bar graph showing energy consumption for current year. The initial implementation and results specified in this section goes one step above to provide users with archived data with 2D interactive visualisation. This would now be the basis for future implementation to develop the planned advanced visualisation, recommendation and the control engines.

VII. FUTURE IMPLEMENTATION The main objective of this research is to develop interactive visualisation software for energy consumption with the planned engines, visualisation, recommendation and control. The initial implementation specified in section VI started off with visualising the energy consumption in the form of 2D

interactive graphs. With the help of end-users (user-centered or participatory design), these graphs can even be better visualised using various visualisation techniques. Later, the recommendation and control engines would be developed as specified in section V.

VIII. CONCLUSION Visualisation of electricity consumption has been gaining popularity with the emergence of android/iOS applications for smart phones and tablets. Although several research institutions and industries targeted towards developing the software-electrical system for VCEC, some of the research questions remain unanswered. The goal of this research is to create a VCEC system model with four main research objectives. They are (a) create usable and useful visualisation of household energy consumption - visualisation engine, (b) create a recommendation engine, (c) create a control engine for visualisation of household energy consumption, and (d) to integrate these 3 subsystems - visualisation, recommendation and control engines. This article is targeted towards the initial step by developing 2D interactive visualisation which will later be enhanced to design an advanced visualisation and control system, which would help the household energy users to reduce domestic energy consumption.

REFERENCES [1] J. K. Dobson and J. A. Griffin, "Conservation effect of immediate

electricity cost. feedback on residential consumption behavior," Proceedings of the 7th ACEEE summer study on energy efficiency in buildings, vol. 2, 1992.

[2] J. Rodgers and L. Bartram, "Exploring ambient and artistic visualization for residential energy use feedback," Visualization and Computer Graphics, IEEE Transactions on, vol. 17, pp. 2489-2497, 2011.

[3] C. AlertMe. (2014, 19 June). AlertMe - Creating Smart Homes. [4] I. EcoDog. (2012). Ecodog - Giving people power over energy. [5] D. Filonik, R. Medland, M. Foth, and M. Rittenbruch, "A customisable

dashboard display for environmental performance visualisations," in Persuasive Technology, ed: Springer, 2013, pp. 51-62.

[6] L. Gamberini, N. Corradi, L. Zamboni, M. Perotti, C. Cadenazzi, S. Mandressi, G. Jacucci, G. Tusa, A. Spagnolli, C. Björkskog, and others, "Saving is fun: designing a persuasive game for power conservation," in Proceedings of the 8th International Conference on Advances in Computer Entertainment Technology, 2011, p. 16.

[7] H. Igaki, H. Seto, M. Fukuda, and M. Nakamura, "Mashing up multiple logs in home network system for promoting energy-saving behavior," in Information and Telecommunication Technologies (APSITT), 2010 8th Asia-Pacific Symposium on, 2010, pp. 1-6.

[8] A. K. Khan, "Monitoring power for the future," Power Engineering Journal, vol. 15, pp. 81-85, 2001.

[9] T. Kim, H. Hong, and B. Magerko, "Designing for persuasion: toward ambient eco-visualization for awareness," in Persuasive technology, ed: Springer, 2010, pp. 106-116.

[10] A. V. Moere, M. Tomitsch, M. Hoinkis, E. Trefz, S. Johansen, and A. Jones, "Comparative feedback in the street: exposing residential energy consumption on house facades," in Human-Computer Interaction-INTERACT 2011, ed: Springer, 2011, pp. 470-488.

[11] E. Costanza, S. D. Ramchurn, and N. R. Jennings, "Understanding domestic energy consumption through interactive visualisation: a field study," in Proceedings of the 2012 ACM Conference on Ubiquitous Computing, 2012, pp. 216-225.

[12] F. Quintal, V. Nisi, N. Nunes, M. Barreto, and L. Pereira, "HomeTree-An Art Inspired Mobile Eco-feedback Visualization," in Advances in Computer Entertainment, ed: Springer, 2012, pp. 545-548.

[13] L. Bartram, J. Rodgers, and K. Muise, "Chasing the negawatt: visualization for Sustainable Living," Computer Graphics and Applications, IEEE, vol. 30, pp. 8-14, 2010.

[14] S. A. Kim, D. Shin, Y. Choe, T. Seibert, and S. P. Walz, "Integrated energy monitoring and visualization system for Smart Green City development: Designing a spatial information integrated energy monitoring model in the context of massive data management on a web based platform," Automation in Construction, vol. 22, pp. 51-59, 2012.

[15] F. Quintal, V. Nisi, N. Nunes, M. Barreto, and L. Pereira, "HomeTree–An Art Inspired Mobile Eco-feedback Visualization," in Advances in Computer Entertainment, ed: Springer, 2012, pp. 545-548.

[16] M. Domínguez, J. J. Fuertes, S. Alonso, M. A. Prada, A. Morán, and P. Barrientos, "Power Monitoring System for University Buildings. Architecture and Advanced Analysis Tools," Energy and Buildings, 2013.

[17] F. Nakazawa, H. Soneda, O. Tsuboi, A. Iwakawa, M. Murakami, M. Matsuda, and N. Nagao, "Smart power strip network and visualization server to motivate energy conservation in office," in Industrial Informatics (INDIN), 2011 9th IEEE International Conference on, 2011, pp. 352-357.

[18] F. Nakazawa, H. Soneda, O. Tsuboi, A. Iwakawa, M. Murakami, M. Matsuda, and N. Nagao, "Contactless power sensor for smart power strip networked to visualize energy conservation," in Networked Sensing Systems (INSS), 2012 Ninth International Conference on, 2012, pp. 1-2.

[19] T. Chiang, S. Natarajan, and I. Walker, "A laboratory test of the efficacy of energy display interface design," Energy and Buildings, vol. 55, pp. 471-480, 2012.

[20] L. Gamberini, A. Spagnolli, N. Corradi, G. Jacucci, G. Tusa, T. Mikkola, L. Zamboni, and E. Hoggan, "Tailoring feedback to users' actions in a persuasive game for household electricity conservation," in

Persuasive Technology. Design for Health and Safety, ed: Springer, 2012, pp. 100-111.

[21] M. Schwanzer and A. Fensel, "Energy consumption information services for smart home inhabitants," in Future Internet-FIS 2010, ed: Springer, 2010, pp. 78-87.

[22] H.-J. Platte, G. Oberjatzas, and W. Vossing. (1988). Remote control device for controlling various functions of one or more appliances.

[23] K. Sakamoto, Y. Iwaji, and T. Endo, "A simplified vector control of position sensorless permanent magnet synchronous motor for electrical household appliances," IEEJ Transactions on Industry Applications, vol. 124, pp. 1133-1140, 2004.

[24] I. Coskun and H. Ardam, "A remote controller for home and office appliances by telephone," Consumer Electronics, IEEE Transactions on, vol. 44, pp. 1291-1297, 1998.

[25] A.-R. Al-Ali and M. Al-Rousan, "Java-based home automation system," Consumer Electronics, IEEE Transactions on, vol. 50, pp. 498-504, 2004.

[26] N. Swamy, O. Kuljaca, and F. L. Lewis, "Internet-based educational control systems lab using NetMeeting," Education, IEEE Transactions on, vol. 45, pp. 145-151, 2002.

[27] K. K. Tan, T. H. Lee, and C. Y. Soh, "Internet-based monitoring of distributed control systems-An undergraduate experiment," Education, IEEE Transactions on, vol. 45, pp. 128-134, 2002.

[28] T. Efergy. (2014). Efergy - Monitor Reduce Save. [29] I. Jacobson, G. Booch, J. Rumbaugh, J. Rumbaugh, and G. Booch, The

unified software development process vol. 1: Addison-Wesley Reading, 1999.

[30] M. J. Muller and S. Kuhn, "Participatory design," Communications of the ACM, vol. 36, pp. 24-28, 1993.

[31] (13 October 2014). Energy in New Zealand [Online]. Available: http://en.wikipedia.org/wiki/Energy_in_New_Zealand