Smart Home Technologies CSE 4392 / CSE 5392 Spring 2006 Manfred

Smart Home Technologies CSE 4392 / CSE 5392 Spring 2006 Manfred Huber [email protected]

Intelligent Environments Environments that use technology to assist inhabitants by automating task components Aimed at improving inhabitants’ experience and task performance NOT: large number of electronic gadgets

Objectives of Intelligent Environments Improve Inhabitant experience: Optimize inhabitant productivity Minimize operating costs Improve comfort Simplify use of technologies Ensure security Enhance accessibility

Requirements for Intelligent Environments Acquire and apply knowledge about tasks that occur in the environment Automate task components that improve efficiency of inhabitant tasks Provide unobtrusive human-machine interfaces Adapt to changes in the environment and of the inhabitants Ensure privacy of the inhabitants

Examples of Intelligent Environments Intelligent Workspaces Automatic note taking Simplified information sharing Optimized climate controls Automated supply ordering

Examples of Intelligent Environments Intelligent Vehicles Location-aware navigation systems Task-specific navigation Traffic-awareness

Examples of Intelligent Environments Smart Homes Optimized climate and light controls Item tracking and automated ordering for food and general use items Automated alarm schedules to match inhabitants’ preferences Control of media systems

Existing Projects Academic Georgia Tech Aware Home MIT Intelligent Room Stanford Interactive Workspaces UC Boulder Adaptive House UTA MavHome Smart Home TCU Smart Home

Existing Projects Industry General Electric Smart Home Microsoft Easy Living Philips Vision of the Future Verizon Connected Family

Georgia Tech Aware Home Perceive and assist occupants Aging in Place (crisis support) Ubiquitous sensing Scene understanding, object recognition Multi-camera, multi-person tracking Context-based activity Smart floor http://www.cc.gatech.edu/fce/ahri/

MIT Intelligent Room Support natural interaction with room Speech-based information access Gesture recognition Movement tracking Context-aware automation http://www.ai.mit.edu/projects/aire/

Stanford Interactive Workspaces Large wall and tabletop interactive displays Scientific visualization Mobile computing devices Computer-supported cooperative work Distributed system architectures http://iwork.stanford.edu/

UC Boulder Adaptive House Infer patterns and predict actions Machine learning for automation HVAC, water heater, lighting control Goals: Reduce occupant manual control Improve energy efficiency http://www.cs.colorado.edu/ mozer/ house/

UTA MavHome Smart Home Learning of inhabitant patterns Learn optimal automation strategies Goals Maximize comfort and productivity Minimize cost Ensure security http://ranger.uta.edu/smarthome/

TCU Smart Home Inhabitant Prediction Smart entertainment control Smart kitchen recipe services Household staff modeling http://personal.tcu.edu/ lburnell/ crescent/crescent.html

General Electric Smart Home Appliance control interfaces Climate control Energy management devices Lighting control Security systems Consumer Electronics Bus (CEBus) http://www.geindustrial.com/cwc/home

Microsoft Easy Living Camera-based person detection and tracking Geometric world modeling for context Multimodal sensing Biometric authentication Distributed systems Ubiquitous computing http://research.microsoft.com/easyliving/

Philips Vision of the Future Less obtrusive technology Technology devices Interactive wallpaper Control wands Intelligent garbage can http://www.design.philips.com/vof

Verizon Connected Family Remote monitoring of the home Entry authentication Integrated, pervasive communications Centralized data management

Challenges in Intelligent Environments Home design and sensor layout Communication and pervasive computing Natural interfaces Management of available data Capture and interpretation of tasks Decision making for automation Robotic control Large-scale integration Inhabitant privacy

Sensors How many and what type? How to interpret sensor data? How to interface with sensors? Are sensors active or passive?

Communications What medium and protocol? How to handle bandwidth limitations? What structure does the communication infrastructure have?

Data Management How to store all the data? What data is stored? How is data distributed to the pervasive computing infrastructure?

Prediction & Decision Making How to extract and represent inhabitants’ task patterns? What patterns should be maintained? How to determine the actions to automate? To what level should tasks be automated?

Automation How are the tasks automated? How are actuators controlled? How is safety ensured?

System Integration How to achieve extensibility? Should the system be centralized or decentralized? How to integrate existing technology components? How to make integration and interface intuitive?

Privacy How to ensure that inhabitant information remains private? What data should be gathered? How should personal data be maintained and used?

Course Topics Sensing Networking Databases Prediction and Data Mining Decision Making Robotics Privacy Issues

Example Scenario Smart kitchen item tracking Sense and monitor items in the kitchen Predict usage patterns Automatically generate shopping lists based on usage patterns Automatically retrieve replacement items