Research and Design Project Ideas from Stephen Intille

See the primary list for a description.

This project will develop a new type of personal, community-strengthening computer device. This device will enrich the lives of residents in traditional housing stock by strengthening their connections with their local neighborhood. The device will facilitate new types of 24-hour communication between people on topics they care about - their family, their neighborhood, their friends, their health and well-being, community safety, local transportation, and educational interests. We call this new device a Universal Personal Interface (UPI). The device is called personal because it will use information collected from each user to automatically customize the device so that it is extremely simple to use. The device is called universal because MIT researchers expect the device to be useful not only in the communities of today; we envision such a device being the primary way that homeowners of the future will interact with their homes and offices of the future. Finally, the device is called an interface because it will promote new types of communication between people, organizations, and other computer devices. To develop the UPI device, we will use commercially proven and widely available wirelessly networked personal digital assistants. These PDAs will be loaded with specialized artificial intelligence software created at the Massachusetts Institute of Technology for the House_n: MIT Home of the Future Project. The software will make it possible for residents of a community to use the PDAs to communicate with each other in new ways to be explored by the research.


This project will develop a new type of personal educational computer interface device that expands the intellectual boundaries of a course beyond the walls of the traditional classroom and university. We will use widely-available wirelessly-networked PDAs to develop specialized software that facilitates merging traditional lecture learning in the classroom with hands-on, experiential learning in the real world. Taking a college course developed to use this device will be 24-hour, lifetime learning experience. It will help keep students connected with the topics in the course many years after the course has been officially completed and simultaneously provide a mechanism by which current students can receive unusual mentoring from MIT alumni of the course and non-MIT professionals. Everyone interested in the topics of the course can learn from databases that contain "personal textbooks" that students use the special interface device to create. We call this device a Universal Personal Interface (UPI) because it will prove useful not only for education today but also as a general, universal interface to the homes and offices of tomorrow. The device will enable and provoke the following activities: - intensive, real-time mentoring between students who have previously taken a course and students currently in a course, - rapid knowledge exchanges between professionals with ideas to contribute but not physical accessibility to the classroom (e.g. the elderly, product designers at Palm, etc.) and course participants, - design of courses where some of the best learning continues for years after the traditional component of the course has officially ended, - assignment of classroom projects that address real-world problems that people encounter every day, - data collection for courses where students are studying the activity of people in the real world, - a merging of the classroom and the home environments, and development of new types of educational interaction in "homes of the future."

Transgenerational, interactive toy
	Level: MEng or Masters with a knack for hardware and low-level software. Some research but mostly technical design and implementation.
	Est. time: 1/2 year for the right person.
	Start: Anytime.
	Description: I have an idea for an inexpensive, but unique, transgenerational and educational toy. There is a small research component, but much of this project would be hacking (and I mean hacking!) some hardware together to demonstrate a concept. When you get it working, you can probably start a company and possibly produce the next Furby sensation.
	Evaluation: Once built, it will be easy to evaluate observing people using it.
	Experience: Background in speech processing and/or linguistics helpful but probably not required.

New sport demonstration
	Level: UROP, MEng. Concept demonstration; no technical research but probably requires some coding to digitally process images.
	Est. time: 1 semester or less, depending upon video editing experience and complexity of the project.
	Start: Anytime.
	Description: Technology enables new types of sport (e.g. cycling, car racing). High-bandwidth home networks and other advanced technologies make entirely new types of sporting entertainment possible. Explore some of these issues in a paper. Select one example and create a video that demonstrates the sport concept and how it might change sport viewership from the home or community. This project probably requires shooting live video and digitally compositing over those shots.
	Evaluation: High-quality video presentation.
	Experience: Video editing, digital compositing, programming (for digital effects).

Digital tag demonstration
	Level: advanced UROP, MEng.
	Est. time: Unknown. Need to talk with research groups with tag technology to determine difficulty of use.
	Start: Anytime.
	Description: Several groups at MIT are developing or using low-cost RF tags for identifying objects. Collaborating with one of these groups, port some of their technology to the House_n prototype space by adding tag detectors to several surfaces in the prototype space and writing interface software to hook into the House_n software infrastructure. Once embedded in the environment, the ability to tag objects can be used to create an application for the space.
	Evaluation: The tags should work reliably and be nicely embedded in the space without a mess of protruding hardware/wires. Documentation should be provided so that other users can write applications using the tags. The demonstration application should convincingly show that the tag system is working.
	Experience: Depends on state of the tagging research. Probably useful to have some hardware experience. Programming expertise in Java and C required to create glue code and write an application that uses tags.

Construction of prototype walls/surfaces
	Level: Any (with appropriate experience). Building; no research.
	Est. time: 1 (busy) month
	Start: IAP 2000.
	Description: Given a design of a prototype room, physically construct the small space consisting of counter-like surfaces and walls with embedded screens. The space may have a raised floor and may require counters and/or cabinets that can be raised and lowered automatically.
	Evaluation: Built to specs.
	Experience: Someone with construction and/or woodshop experience and a knack for building things that look professional and who can follow this through from buying supplies to final coats of paint.

Wearable setup
	Level: Any. Computer hardware/software hacking.
	Est. time: Several weeks.
	Start: Anytime.
	Description: Following the specs laid out by the ML Wearables group, build a wearable computer. Alternatively, justify why buying a commercial system is the best thing to do. Do all the groundwork ordering parts, etc. Implement communication with the room environment (preferably wireless but tethered ok). Wear it out in public to make sure it works robustly and then package it up so someone else can develop software for it, if needed.
	Evaluation: It works and is well-documented so that someone else can use it.
	Experience: Interest in this sort of thing.

Web: social anthropology resources
	Level: UROP
	Est. time: 1/2 semester
	Start: Anytime.
	Description: Compile materials online on social anthropology. These resources should be organized so that someone unfamiliar with the field can learn the basics. In particular, the site should contain information on qualitative and quantitative evaluation methods and an example of how to perform observations of people using devices and test some hypothesis. The goal is not just to compile information but instead to distill it down so that it is useful to engineers and architects.
	Evaluation: Clean, easy-to-navigate set of comprehensive resources. A non-anthropologist should be able to read the site and have a general idea of how to perform a social anthropology study (and especially what pitfalls to avoid!).
	Experience: Anthropology experience would be nice, but it's not required; attention to detail; a desire to quantify when possible.

Web: predictions for the year 2000
	Level: UROP
	Est. time: severa; weeks
	Start: Anytime.
	Description: Compile materials online on predictions that people made about technology and how we live for the year 2000. Compare these to what we have today.
	Evaluation: Clean, easy-to-navigate set of comprehensive resources.
	Experience: Good library research skills; attention to detail.

Observing and interviewing nuclear families in the home
	Level: Any. Social anthropology project.
	Est. time: 1-2 months.
	Start: Any.
	Description: Obtain permission to observe three unrelated, "typical" families in their own homes for extended periods of time. Ideally the families should have children in the home. Record your observations and use them to propose several new ideas for devices (or redesigns of existing devices) to solve some observed problem(s).
	Evaluation: A paper detailing observations and justifying the device designs.
	Experience: Interest in social anthropology; interview skills; patience.

Homes of India and China
	Level: Any. Library research.
	Est. time: 3 weeks.
	Start: Any.
	Description: Compile resources on "typical" homes in India and China. Document the differences between these spaces and American homes.
	Evaluation: A paper comparing and contrasting interaction in these homes and speculate how the home of the future in India and China should differ from the home of the future in the US and describe why. In particular focus on sustainability issues. How can the home of the future be designed in an environmentally friendly way given the explosion in the number of people who will need housing in the next century?
	Experience: Architectural background desirable but not required.

Bolo game replacement
	Level: Any. Software programming.
	Est. time: Unknown ... substantial work.
	Start: Any.
	Description: Video games are excellent platforms for studying action recognition algorithms. See Inducting Indy for one example that used the Mac game Bolo. Unfortunately, Bolo is a Mac game, making it difficult to use for research in labs dominated by PC machines. Write a new networked game, based on Bolo or not, that could be used for action recognition research (one possibility is to start with the partially-developed version of Bolo for the PC).
	Evaluation: The game must be very robust so that it can be widely distributed and good (so that people will actually play it without being paid!).
	Experience: Extensive Windows programming experience; game programming experience; interest in computational perception.

Designing an energy efficient kitchen using technology and integrated appliances
	Level: Any. Library research, architectural design, and product design.
	Est. time: Several months.
	Start: Anytime.
	Description: Perform library research on the efficiency of modern kitchen appliances. Use this background research to design a new kitchen where some or all appliances are somehow integrated with one another to provide greater energy (and water usage) efficiency. Carefully model your design and justify your decisions. Your technical assumptions should be reasonable given current technology.
	Evaluation: Generate computational models of the kitchen that includes animated demonstrations of moving components. The design decision should be strongly supported in an accompanying paper that quantitatively argues for the kitchen's improved efficiency.
	Experience: Architecture/design background or product development experience with the ability to quickly pickup technical ideas; visual rendering of models.

Report on self-sufficient homes
	Level: UROP. Library research.
	Est. time: 2 months.
	Start: Any.
	Description: Investigate current and past examples of attempts to build self-sufficient environments and characterize the issues that have been encountered and the problems that remain to be overcome. Examples of such environments would include the Biosphere, small test environments build by NASA, etc. The findings should be used to propose ways of potentially improving the self-sufficiency of the home of the future.
	Evaluation: The research must be published online and is expected to become the most comprehensive website for information on self-sufficient homes online.
	Experience: Interest in sustainability; attention to detail; familiarization with web publishing.

Eliminating urban eyesores: parking meters, road lights, mailboxes, floodlights, etc.
	Level: Any.
	Est. time: 2 months.
	Start: Any.
	Description: Propose a set of solutions (hopefully not independent of one another) to eliminate urban eyesores from communities. Document solutions that have been proposed by others.
	Evaluation: Paper publishable on the web that includes some detailed design sketches of the proposed solutions.
	Experience: Interest in architecture and/or product development; ability to get your ideas into a visual format for final report; familiarization with web publishing.

"Electronic ink surfaces:" surfaces that change color and their tradeoffs
	Level: Any. Depends on scope.
	Est. time: Unknown.
	Start: Any.
	Description: Investigate the impact that surfaces that change color (and visual texture) using future technology analogous to electronic-ink (e.g. eink) will have on interior spaces in the home. The House_n prototype space might be used with video projection on surfaces to simulate a possible application or effect. Alternatively, video editing and digital compositing might be used to simulate some application. Another option for a a technically-minded student with a hardware/physics background might implement a small surface embedded into a counter top.
	Evaluation: Depends on project scope.
	Experience: Depends on project scope.

Design an interactive home for the average world citizen
	Level: Any. Design.
	Est. time: .5 - 1 semester
	Start: Any.
	Description: The world's limited resources simply cannot support the type of wasteful, opulent homes found in the US. The average world citizen current has very little money, education, and available space. The real home of the future must provide a livable environment for the average citizen while consuming very little space and natural resources. Design such an environment, using the clever integration of architecture and technology to provide a small, adaptive space that is both environmentally friendly and a desirable, satisfying living environment.
	Evaluation: Interior and exterior computational or physical models of the home design that include detailed descriptions of how each interface device would be used and narratives describing how people would use the space.
	Experience: Architectural CAD skills.

Football web interface data collection
	Level: UROP.
	Est. time: 1 semester.
	Start: Any.
	Description: A Java web interface has been developed that allows football fans online to label action in football plays. See the current interface and the Computers Watching Football homepage for more detail. The data collected would be used in future research on developing algorithms to automatically recognize actions in football plays. Although the online labeling interface exists, it hasn't yet been advertised well and has not yet been refined to actually collect data. This project involves making any minor fixes to the Java code required to make this a usable interface and building a community of online users who will provide a large database of labeled football actions.
	Evaluation: The database should be packaged up in a form that is easy for researchers to use. It should include entries from thousands of football experts.
	Experience: Java programming for the web; knowledge of football; interest in football; patience to handle inquiries from online users; attention to detail.

Football activity recognition
	Level: Masters. Research.
	Est. time: Unknown.
	Start: Any.
	Description: This research involves exploring ways to improve prior research into understanding multi-agent action for recognition of activity in football games (see the Computers Watching Football homepage).
	Evaluation: Publishable quality work evaluated on the football data sets.
	Experience: Course in image processing and AI; programming experience.

Visual dishwashing
	Level: Any. Depends on scope.
	Est. time: Unknown.
	Start: Any.
	Description: Design a better dishwasher that uses computer vision technology (and possibly other technical and design innovations). Implement a demonstration of one of the algorithms (e.g. if it is detecting how dirty a dish or water may be, detecting if people are near the dishwasher, etc.).
	Evaluation: This project should result in a report describing current dishwasher technology and a demonstration showing one perceptual sensing algorithm that might be incorporated into a "dishwasher of the future" to improve some aspect of the device's performance.
	Experience: Java/C programming; image processing or AI course.

Context-aware toilet
	Level: Unknown.
	Est. time: Unknown.
	Start: Any.
	Description: Using sensing technology, is it possible to design a better toilet? Consider what "better" means and make a proposal. Be sure to base any design on background research on the design of toilets, both in the US and in other parts of the world.
	Evaluation: Depends on scope.
	Experience: Depends on scope.

"Plug 'n play" architectural panels
	Level: 1 semester.
	Est. time: Description.
	Start: Any.
	Description: The goal of this project is to design (on paper or computationally) the ultimate architectural "plug 'n play" panel. By connecting these panels together, it should be possible to create a livable environment that has the infrastructure to become computationally enhanced. Issues to consider are the shapes of the panels, how they will interlock, how they provide insulation, how they provide display capabilities, how they permit wiring to run through the home, how they can be easily replaced, what sort of sensors each panel should contain, how much they might eventually cost (with economies of scale), etc. Although a few specialty panels will no doubt be required, the focus should be on the "core" panel.
	Experience: The panel should be described in great detail down to the last material used. A physical or computational model should be constructed. Also shown should be a few representative forms the panels are capable of creating. A report should describe the various tradeoffs required when designing the panel.
	Experience: Ability to integrate ideas from several fields; ability to construct physical or computational models; attention to detail.

Building a "plug 'n play" architectural panel prototype
	Level: Masters, Ph.D.
	Est. time: 1+ years.
	Start: Any.
	Description: Build a physical model of a panel design (see "'plug n play' architectural panels" project).
	Evaluation: The panel should be operational (if a bit more bulky than it might be in real life). At least two panels will probably be needed to make a convincing case for the design.
	Experience: Extensive experience working with materials and electronics and building physical models of devices.

Virtual fridge
	Level: Any. Interface design and programming.
	Est. time: 1.5 - 2 semesters.
	Start: Any.
	Description: Assume that kitchen supplies, including the food and leftovers in the fridge, are all stored via an automatic shuttling mechanism. A user of the kitchen simply puts a bit of leftovers in a container, inserts the container into the shuttling mechanism, and away it goes to be kept cool. One problem with this system is that a person needs a way to figure out what is stored in the fridge. A person might even want to know information about environmental conditions in the fridge (e.g. moldy and smelly?). Design an interface for the House_n prototype space that allows users to view and select items in a fridge which is not physically accessible (the interface might use iconic representations of items or actual video images of objects that have been organized/manipulated in some way).
	Evaluation: Working interface that reliably demonstrates the idea for a fabricated test fridge scenario.
	Experience: Java programming.

Mechanism for product transport: integrated core kitchen unit
	Level: Masters, Ph.D. Research/design.
	Est. time: At least 1 year.
	Start: Any.
	Description: Design and build a mechanical system for transporting specially-designed containers that store goods from the home. The system should meet constraints laid out by the House_n research into ways that people are likely to use their homes in the future.
	Evaluation: Development of a working system. Attributes of most significance are likely to be speed, the ability for non-experts to perform maintenance, and the capacity to carry heavy objects at low cost.
	Experience: Mechanical engineering with experience building robotic devices; some programming to write software control.

The compact home: a compact, efficient, rich interactive environment
	Level: Any. Design.
	Est. time: 1 semester.
	Start: Any.
	Description: There are many reasons to design very compact homes that provide rich living experiences. They may be more portable and require less resources to build and maintain. When they do need to be decommissioned, they leave less waist. Organizations like military units, disaster relief providers, and construction crews need to rapidly deploy non-permanent living spaces. Perhaps the most important reason to design compact homes is the rapidly growing world population that requires (and will require even more) housing .. often with low cost and low space requirements. Design a compact living environment that uses interactive technologies to make the space adaptable, comfortable, and enriching. Model the space and then document how people would live in the space and how many resources typical use of the space would require. Compare the amount of materials required to construct this space on a large scale (hundreds of millions) with the amount of material that would be required to construct a typical American home. How small can the livable space be made without encroaching on quality of life?
	Evaluation: Detailed computational models of the interior and physical and/or computational models of the exterior with a paper justifying various design decisions.
	Experience: Architecture; comfort with technology.

Design a methodology for designing communities of compact homes
	Level: Any. Design.
	Est. time: .5 semester.
	Start: Any.
	Description: Assume the existence of some type of very compact home (see "The compact home: a compact, efficient, rich interactive environment" project). Establish a methodology that can be used to rapidly construct a community. The goal is to identify criteria that will help to rapidly create a real community that doesn't feel like a temporary establishment. The methodology should account for different topologies and, to some extent, different communities of people who might form such an instant community. The implications of using digital communities and wearable computers in addition to more traditional urban planning criteria should be considered.
	Evaluation: Detailed computational models of the interior and physical and/or computational models of the exterior with a paper justifying various design decisions.
	Experience: Urban planning; comfort with technology.

The virtual meal - interface development
	Level: Any. Software programming and interface research.
	Est. time: 7+ months.
	Start: After prototype space is constructed (mid-January) and the "virtual meal" project above is completed.
	Description: Design an interface to address the issues discovered with infinite bandwidth communication.
	Evaluation: Test the interface with real subjects (to do this it must be very intuitive and robust) simultaneously using the prototype space and the satellite space.
	Experience: Java programming; interest in virtual communication; communities.

Resources on IEs
	Level: UROP. Internet research and compilation.
	Est. time: 2 months.
	Start: Anytime.
	Description: Compile the best online and offline resource on the design of interactive environments. The main body of the site should be an annotated list of websites and published articles highlighting the interesting points or ideas of each article.
	Evaluation: Clean, easy-to-navigate set of comprehensive resources. A non-computer scientist should be able to read the site and have a general idea of what the hard and easy problems are.
	Experience: Interest in technical and design aspects of interactive environments; familiarization with online publishing; attention to detail.

Homes of the Future of the Past: technology marketing
	Level: MBA.
	Est. time: 1 semester.
	Start: Anytime.
	Description: There are some good ideas in some of the homes of the future of the past. Why didn't they catch on? What marketing ideas could be brought to bear on this problem? Does technology offer any solutions?
	Evaluation: Unknown.
	Experience: Business marketing.

Enabling devices
	Level: Any.
	Est. time: Depends on number of devices connected.
	Start: Anytime.
	Description: Write drivers or interface code to allow various electronic devices to be used in an interactive space. Examples of devices might include ID scanners, card scanners, tablets, projectors, mounted cameras, slide projectors, etc. For the most part this will involve writing wrapper code to interface with the room's Java platform.
	Evaluation: Working interface code.
	Experience: Java programming; software device-level programming.

Visualization connections between home devices, technologies, needs, and values
	Level: Any.
	Est. time: 2 months.
	Start: Anytime.
	Description: Create a graphical spring model Java interface showing the connections and strengths between architectural concepts, values people have for their homes, and particular devices and technologies. The model can be used to visualize what a change in a value might impact on the technology in the home.
	Evaluation: Easy-to-use visualization tool. It should be straightforward to add and remove connections either through the interface or through a text file that gets loaded.
	Experience: Java programming; interest in homes of the future. This is a good project for someone who wants to learn Java.

Seamless surfaces
	Level: Any.
	Est. time: Unknown.
	Start: Anytime.
	Description: Model a kitchen/living room space (analogous to the prototype space) where all the problematic seams between surfaces (e.g. where dirt gets caught) have been eliminated. Propose a construction method and use of materials. Now do the same for a bathroom space. Discuss why seamless bathrooms haven't caught on.
	Evaluation: Completed computational and/or physical models with justification of design decisions documented in a paper that includes relevant background research.
	Experience: Architecture; construction; materials.

Implementing wired speech recognition software
	Level: UROP. Software programming.
	Est. time: 3-6 months.
	Start: Anytime.
	Description: Using the Dragon or IBM speech recognition development packages (for NaturallySpeaking or ViaVoice, respectively), create wrapper code that hooks into the prototype space Java architecture so that keywords can be recognized when space occupants are wearing lapel microphones. Use a setup with wireless microphones, and design the code so that a list of expected words can be provided to the speech code to limit the context and improve recognition.
	Evaluation: Working system that can easily be used by other application developers.
	Experience: Java/C++ programming.

Simulating people for urban planning
	Level: Ph.D.
	Est. time: 2+ years to do it right.
	Start: Anytime.
	Description: Study the work that has been done at MIT on simulating driver behavior using MITSIM and other "people simulators." Develop behavior models for people and use those models for simulating people for an urban plan evaluator.
	Evaluation: Publishable work in urban planning and other fields.
	Experience: Interest in cognitive science and architecture; optimization.

Visualizing values
	Level: Unknown.
	Est. time: Unknown..
	Start: Anytime.
	Description: Unfortunately, people often based their home buying on surface criteria (e.g. size of master bathrooms) instead of usability criteria (e.g. ease of transferring groceries from the car to the kitchen). Develop a system that allows a prospective purchaser of some home of the future to determine if the home will satisfy fundamental user values.
	Evaluation: Unknown, but some mechanism is required.
	Experience: probably Java/C programming; virtual reality (possibly); marketing.

Implement a virtual projector algorithm for curved surfaces
	Level: Unknown.
	Est. time: Unknown.
	Start: After previous project works.
	Description: Tiling multiple projectors on curved surfaces.
	Evaluation: It works -- meaning calibration looks good to human observers -- and the types of errors people can see on a test set have been characterized. A person unfamiliar with the system should be able to use it to calibrate video projectors. The type of curved surfaces the system can and can't handle should be specified.
	Experience: Programming; image processing; projective geometry; possibly learning.

Robust, wireless communication
	Level: UROP, MEng. Software programming with devices.
	Est. time: 1 semester.
	Start: Anytime.
	Description: Implement a wireless communication system for message passing (including possibly images) between a wearable computer and a network of desktop PCs. Use commercial hardware and write the glue code.
	Evaluation: It works ... reliably ... for hours at a time ... and is easy to startup and install.
	Experience: Java/C programming; WinNT familiarity.

Web: Home networking
	Level: UROP. Library research.
	Est. time: 1.5 months.
	Start: Anytime.
	Description: Compile a comprehensive website (consistent with House_n format) on home networking systems. The goal is not just to compile information but instead to distill it down so that it is useful to engineers and architects. The website should include descriptions of how the enabling technology works for the different home networking options and the physical limitations of each method.
	Evaluation: Clean, easy-to-navigate set of comprehensive resources. The format should make it easy to add updates as new systems come to market and appear in research literature.
	Experience: Interest in wireless communication desirable; familiarization with online publishing; attention to detail.

Integrate wearable display interface and room display
	Level: UROP,MEng. Software programming.
	Est. time: 1 semester
	Start: Anytime.
	Description: Demonstrate an application that integrates a wearable computer display (e.g. using MicroOptical glasses) and a display projected onto a surface in the House_n prototype environment to create some type of computer interface for interaction in the home.
	Evaluation: A user should be able to put on the wearable computer and interact with the system without any guidance from the system designer. Application should permit a new type of interaction or service in the home.
	Experience: Java/C programming experience.

Visualization space
	Level: Any.
	Est. time: Unknown.
	Start: Anytime.
	Description: The prototype House_n space has been designed so that video can be recorded of people interacting in the space and then video compositing can be used to demonstrate new architectural principles. This project involves creating one such simulation and developing software tools to facilitate this process.
	Evaluation: One several minute "visualization" and documentation of the process used to create the video.
	Experience: Video editing; software programming; computer graphics; possibly projective geometry.

"Sleeping box"
	Level: Any.
	Est. time: Unknown.
	Start: Anytime.
	Description: Design a "sleeping box" (think of a better name!). This is a small space where someone can retire for the night, perhaps in a hotel of the future. It should provide a comfortable sleeping space and interactive entertainment, but all in a small package not much bigger than a casket (gruesome, I know). Figure out what needs to be done to make this type of sleeping space non-claustrophobic and comforting.
	Evaluation: Depends on scope of the project. If it is modeled, several applications should be simulated. If a working prototype is constructed, some people should sleep in it a few nights and their observations should be recorded.
	Experience: Depends on scope.

Build an adjustable workstation table
	Level: Any. Building.
	Est. time: Description.
	Start: Anytime.
	Description: Design a comfortable and ergonomic table. Start by studying the Levity. Now improve the design and build an actual table.
	Evaluation: The table should be usable and their should be some documentation to help someone if they want to build another.
	Experience: Building things with wood or other materials; keen powers of observation; architectural design.

Designing an ergonomic self-monitoring adjustable workstation table
	Level: Any. Research and software development.
	Est. time: 1 semester.
	Start: Anytime.
	Description: Use computer vision and/or a sensor pad to develop a workstation table that can warn a user when he/she is doing something that the user is trying not to do. For example, computer vision could be used to tell someone when they might be bending the wrist too much when using the mouse or when the person is slouching. This might be achieved by showing the system examples of undesirable and desirable positions so the system can classify new images captured periodically from a nearby video camera. Ideally this project would be integrated with the workstation table described in the "build an adjustable workstation table" project.
	Evaluation: The table should be usable and their should be some documentation to help someone if they want to build another.
	Experience: Pattern recognition course; Java/C programming; possibly learning algorithms.

Observing behavior in the home
	Level: Any.
	Est. time: 1 semester plus several 48 hour marathon sessions
	Start: Anytime.
	Description: The goal of this project would be to create a unique and comprehensive video dataset of activity in a typical home. The goal would be to find 2-3 families who will allow their every move to be recorded in their homes for scientific study. Sensors (primarily cameras and microphones but potentially others) would then be temporarily installed in a home. The activity in the home would then be recorded in such a way so that all interesting interaction is captured for a 24-48 hour period. This project is a logistical challenge, to put it mildly. Problems (among others) are (1) finding suitable subjects, (2) putting sensors in homes temporarily that cover every space in the home, (3) figuring out a way to record/store the tremendous amount of video that will be recorded, (4) keeping the cost of equipment needed (for purchase or rental) realistic, (5) developing a format used for tagging the video so that interesting activities are recorded, and others.
	Evaluation: The hope is that this database -- despite consisting of only a few examples -- could be used to help researchers identify some patterns of behavior and create models of activity. Algorithms for detecting activity could be tested using the dataset. The project's success would be judged on the usability of the database as a research tool.
	Experience: Comfort with setting up computer equipment and video equipment; ability to work with people outside of MIT; high level of maturity.

Bottle laws and resource consumption
	Level: Any. Library research.
	Est. time: 4 weeks.
	Start: Anytime.
	Description: Explore the impact that bottle laws have had on the amount of material recycled, paying special attention to Massachusetts but considering other states. The bottle law has been in the news lately because it may get repealed (9/99). Find the facts on how much impact the $.05 charge has had on consumer behavior. Further, try to evaluate what the law's impact has been on cleanliness in communities. What light (if any) does this attempt to influence consumer behavior shed on the design of other interfaces in the home to reduce consumption? Consider the interesting phenomena of "secondary" recycling, where people who don't recycle end up recycling when enterprising individuals pull cans and bottles from the trash. Can this idea be extended to other home resource issues? The project may require doing some interviews with recycling personnel, city cleaning personnel, and typical bottle users.
	Evaluation: Publish your results online for others to use as reference.
	Experience: Familiarization with web publishing.

Redesigning the window
	Level: Any. Research, library research, building, programming.
	Est. time: 1-2 semesters, depending on the scope.
	Start: Anytime.
	Description: Windows are due for a redesign using advanced materials and technology. Design a new window unit for a standard home meeting a new set of criteria for this this important information appliance.
	Evaluation: Construct a functional prototype and write a paper evaluating characteristics (e.g. energy use, interface usability) of the proposed device.
	Experience: Building, architecture, programming.

Install the MIMEO device
	Level: UROP, MEng. Programming.
	Est. time: .25 semester, depending on the scope.
	Start: Anytime.
	Description: Install a MIMEO whiteboard device in the prototype space and write software to interface the device with the prototype room's Java system. Design some application using the MIMEO that demonstrates a new idea related to education in the home.
	Evaluation: Construct a functional prototype of your educational scenario. Code should be well-documented and easy to use so someone unfamiliar with it can use the MIMIO as an input device for other applications.
	Experience: Java/C programming; possibly driver hacking.

Robotic sweeper
	Level: MEng, Masters. Hardware engineering, software engineering.
	Est. time: 1 semester to build, longer to program control code.
	Start: Anytime.
	Description: Design a very small, very fast "robotic sweeper." The sweeper will use information provided by the prototype House_n sensors for navigation. Assume that the robot can be controlled (to some extent) from an external camera and that roaming objects (e.g. people) will be in predictable locations. Therefore, the design should focus on quickness and robustness given a particular trajectory, not navigation. The sweeper would demonstrate the ability to clean something up. For example, it might have a very small vacuum cleaner inside and have to clean up a small mess of dropped punched holes. Alternatively, it might dispense some fluid, such as Windex, on a hard surface and use a wiper device to polish a surface clean. In either case, the robot should be designed to be small and fast. The robot would be used in the following scenario. When people are in a kitchen space, the robot will be in a compartment in the wall. However, when people leave a mess (detected using computer vision techniques, most likely, but your robot doesn't need to detect the mess), the robotic sweeper will come out (when the people leave or go away from the mess) to clean up the area. It may take it more than one trip, which may be required to keep the device light enough and fast enough to literally zip around the space. Ideally, the physical design of the sweeper will be integrated with the design of the floors and walls to allow cleaning the "hard to get" spaces at the wall/floor connection (see "Seamless Surfaces" project).
	Evaluation: A working, functional prototype made of off-the-shelf components or that can be reconstructed if duplicates must be made.
	Experience: Robotics, mechanical engineering.

Robotic container design
	Level: Any. Mechanical engineering, design, usability research.
	Est. time: 1 semester
	Start: Anytime.
	Description: Design a system of containers for storing food in the home. These containers should be designed so that they can be moved by automated systems quickly and efficiently in a "core" kitchen appliance unit of a home. Containers should be designed to carry objects found in the normal kitchen including individual pieces of silverware, individual plates and cups, large cartons (e.g. milk, detergent), unpackaged fruits and veggies, and leftover food. The containers should make it possible to transport items directly from and into appliances such as washing machines, refrigerators, microwaves, and ovens. In addition, the containers should allow for efficient storage. Finally, the containers should be designed to facilitate automatic delivery from the core unit to the home space. The containers should look appealing when/if they every appear in the home space. Specify the design of the "portal" space where containers move from the core to the home space and back. Describe how people would use the containers. The containers must not depend upon any technology that is not currently deployable in order to function.
	Evaluation: All the containers should be modeled digitally. Clear justification of all design decisions must be reported. At least one real container must be constructed.
	Experience: Mechanical engineering; product design; keen powers of observation.

Robotic container dispenser/collector
	Level: MEng, Masters. Mechanical engineering, software engineering.
	Est. time: 1 year
	Start: Anytime.
	Description: See the "robotic container design" project. Design a mechanical engineering system that can move these containers from the prototype home space to the core space. Only a small number of small objects (e.g. a few pieces of silverware and plates and cups) need to be handled. The goal is to simulate how a full-blown system would actually work. Therefore, the mover must be extremely fast to facilitate natural interaction. It is ok to assume that this transport unit might pass off the objects/containers to a slower transport device. The focus here is to have a real device that people in the prototype can interact with at real-time rates so that other students can experiment with interface design.
	Evaluation: A working, functional prototype that can be controlled via PC software.
	Experience: Robotics, mechanical engineering.

Assessing user awareness of resource consumption
	Level: Any. Research.
	Est. time: .5 semester.
	Start: Any.
	Description: Do some background research compiling statistics of resource consumption by the average American using different appliances. Then design a survey (getting input from researchers experienced doing such things) to evaluate how well people understand how much energy and water they consume and which devices they use consume the greatest percentages.
	Evaluation: Publishable results from survey, if survey is designed well.
	Experience: Psychology questionnaires.

Mobile "Stringer" reporting wearable
	Level: UROP, MEng.
	Est. time: 1.5-2 semesters
	Start: Any, once wearable computer is available.
	Description: Look at the Silver Stringers website. Design an interface for a wearable computer that would assist a senior citizen who is acting as a reporter in the community. What does the physical interface used on the wearable (as well as the digital interface) need to look like and how does it work? Think about how this interface must interact with devices in the home.
	Evaluation: Produce a simple, but working, wearable interface that is very easy to use. Obtain feedback from at least two members of one of the Silver Stringer communities. In addition, provide a report modeling a better user interface for the wearable (if a new physical interface must be designed) and detailed descriptions of how the wearable connects to devices in the home of the future.
	Experience: Java programming; interview skills.

Community and senior reporting
	Level: UROP. Speculative paper.
	Est. time: 4 weeks.
	Start: Any.
	Description: Assume that the Internet has made it possible for non-manufacturing businesses and residential spaces to be highly intermixed (i.e. housing and offices are in the same building, and possibly even on the same floor). Assume the residents benefit from having business nearby by sharing resources. What is the benefit to businesses? Think about this. Consider one benefit might be achieved by providing tools to senior citizens to help them report on businesses for business and to help establish a sense of community. Propose a high-level plan for how this might work. Run the idea by seniors and business executives and report on their feedback.
	Evaluation: A well-reasoned report that document the idea and feedback obtained.
	Experience: Interview skills.

Reinventing the lecture
	Level: Masters.
	Est. time: 2 years.
	Start: Any.
	Description: Take a typical large-lecture course at MIT. Develop a new type of "interactive lecture experience" for the course using the standard course material and assignments but using advanced technologies to make the lectures and recitations more engaging.
	Evaluation: Best would be some comparison between the scores of the modified lecture and the original lecture on a final exam, but this would be tricky to do in a meaningful way.
	Experience: Software programming, interest in teaching, ability to work with professors, possibly video and sound editing.

Irradicating the iron
	Level: Masters+.
	Est. time: Unknown.
	Start: Any.
	Description: For people who don't care about "pressed" clothes, only non-wrinkled clothes, the design of a new type of dryer can be used to eliminate the iron. You would assume that all clothes come with a tiny ID tag that can be used to access the exact make of the shirt (i.e. material properties). Design a new type of dryer that will pull clothes from the main drying bin at just the right time to facilitate wrinkle-free hang drying. Determine a way for the clothes to all be held at the perfect temperature/dampness until the entire load is done (or to actually hang the clothes as they finish!).
	Evaluation: Model this device digitally and back up all your technological assumptions. Possibly build a prototype if a MechE.
	Experience: MechE, mechanical product design.

Food consumption
	Level: UROP.
	Est. time: 4 weeks.
	Start: Any.
	Description: Perform library research to categorize the types and volumes of food discarded each year by people in homes. Is the food spoiled? Why? Unwanted leftovers? Use the research to propose some ideas of how the amount of waste could be reduced using architectural design and technology.
	Evaluation: Report published online.
	Experience: Attention to detail.

Optimizing vehicle use in the community
	Level: Unknown.
	Est. time: Unknown.
	Start: Unknown.
	Description: Develop a system that observes activity of people over time and uses information about patterns of activity, what they need, where they live, what's nearby, the efficiency of their vehicle, time constraints, and community congestion/activity patterns to develop a scheduling algorithm and user interface for the home that helps someone cluster errands that are run in the family vehicle. The goal would be to provide more free time, reduce use of gas and pollution emission, and simplify the shopping tasks.
	Evaluation: Unknown.
	Experience: Unknown.

Containers and the car
	Level: Unknown.
	Est. time: Unknown.
	Start: After container project.
	Description: Take the set of containers developed for transporting goods in the home (see project above). Describe how the containers impact the design of parts of compact cars (e.g. the trunk). Design a system for docking the car's trunk with the home's container shuttling mechanism.
	Evaluation: Unknown.
	Experience: Mechanical engineering, design.

Environmental displays for driving
	Level: Unknown.
	Est. time: Unknown.
	Start: Unknown.
	Description: There are probably interface issues with how you might convey information about your car and energy use to someone in the home. Analogous to the health care information display problem.
	Evaluation: Unknown.
	Experience: Unknown.

3D printing thought experiment
	Level: UROP.
	Est. time: 3 weeks.
	Start: Any.
	Description: Assume color 3D printing technology is as ubiquitous, cheap, and fast as color inject printers and printing are today. What will people use the 3D capability for? Try to predict some applications -- especially applications that are not work related. Given foreseeable technologies, what technological limitations do you predict?
	Evaluation: A quick report listing your ideas, published online.
	Experience: Visionary thinking.

Community components
	Level: Unknown.
	Est. time: .5 semester.
	Start: Anytime.
	Description: House_n researchers are investigating modular components used to construct homes. Consider how you might develop "community components" that allow a community of the future to expand gradually. Mechanisms must be developed that allow cables, shuttle mechanisms, power sources, etc. to be slowly expanded as a community grows. I have some ideas on this.
	Evaluation: Unknown.
	Experience: Some urban planning helpful.

Integrating corporate and residential spaces using active badges
	Level: UROP.
	Est. time: 2 months.
	Start: Any.
	Description: Review the ideas developed by House_n researchers on the positive and negative reasons to very tightly integrate commercial and residential spaces in the communities of the future. One major factor prohibiting architectural integration is the concerns of businesses about protecting their IP and maintaining a sense of a corporate identity. Propose a coherent plan showing how technology such as active badges and buildings where all walls can be interactive displays can solve these problems.
	Evaluation: Presentation of one coherent architectural plan for a small community including sketches or renderings of some spaces, detailed arguments for how the technology will work, and discussion of the pros and cons of the proposed strategy. Published no the web.
	Experience: Architecture and/or urban planning; ability to produce sketches/drawings/renderings.

Movies of the Future
	Level: UROP.
	Est. time: 2 weeks.
	Start: Anytime.
	Description: Compile an annotated bibliography of movies and television shows that are in some way related to "home of the future" topics. There is a start at: http://architecture.mit.edu/house_n/web/resources/links/movies.html.
	Evaluation: Completed and comprehensive.
	Experience: Movie buff would help.

Gadget guru
	Level: UROP/MEng
	Est. time: Unknown.
	Start: Anytime.
	Description: In this project, you will contact companies who make interesting hardware devices, try to convince them to donate units, and then work on writing glue code so that other programmers can access the devices from Java code. .
	Evaluation: You will need to get one or more devices working reliably so others can use them and provide short writeups on the positives and negative characteristics of each device.
	Experience: Comfort with driver programming helpful; able to interact with representatives from companies; an interest in the latest gadgets; Java/C programming.

Website for ranking people activities in the home
	Level: UROP
	Est. time: Several weeks.
	Start: Anytime.
	Description: Develop a website that allows Internet users to help House_n rank the importance of person activities in the homes of the present and future. For example, for a long list of actions that people do in the home (cooking on the stove, cutting hair, checking the fridge, waking up, cleaning floors, etc.) visitors to the website would be able to easily specify things like whether they want to increase or decrease the amount of time they spend doing each activity, whether they want to increase or decrease the quality of the activity, etc. You would also do some work testing the site to make sure it is solid and then trying to attract people to the site so we can start collecting data.
	Evaluation: Works very robustly.
	Experience: Comfortable with web programming so that you can have a database that is manipulated using web forms.

Transgenerational educational messages
	Level: UROP/MEng
	Est. time: Depends on scope.
	Start: Anytime.
	Description: The idea for this project is that we will have PDA type devices that will receive educational messages. You will pick a topic of interest and develop software that automatically collects factoids from the web that could be sent to a person who wanted little "newsbytes" about the topic -- kind of like one of those "word a day" calendars. Next you will do some simple processing on these phrases and try to come up with a system that can parse the factoids into two versions: one for adults and one for children. The same information will be conveyed, but the way it is conveyed will need to change. The more that can be done automatically, the better.
	Evaluation: Demonstrate some compelling examples for a small topic area.
	Experience: Java programming; understanding of parsing helpful; interest in children and learning.

Robotic soccer
	Level: Any.
	Est. time: Depends on scope, but at least 1 semester just to get started.
	Start: Anytime.
	Description: There is a well-organized robotic soccer competition. You would form a team for the simulator division, where you write computer code to control your sim team. You will need to spend many hours working on programming, and you will need to help recruit a team of MIT students to help you. A good result of this project would a self-running MIT club to continually develop a robotic soccer team. To get your simulated team to do anything interesting is difficult and there are many underlying research issues that could be addressed.
	Evaluation: You can enter the yearly robotic soccer competition and see how your "team" does against teams from other schools around the world.
	Experience: Programming; like writing efficient code; interest in AI.

Setup a full-blown web server
	Level: Any. Strong interest in web/systems hacking.
	Est. time: Depends on experience.
	Start: Anytime.
	Description: House_n could use it's own database backed server. Use PCs (either with NT server or with Linux) and a good database (e.g. Oracle, which MIT can get for free) get a stand-alone server running that can handle very high volume.
	Evaluation: It works and other people can use it as a Java server.
	Experience: You should have done this before or have a true love for UNIX/systems stuff.

Simulating people movement and interaction Level: Masters. Research. Est. time: 1-2 semesters or more. Start: Any. Description: Given a topology that includes a set of urban forms (e.g. roads, buildings of particular types, public transportation systems), develop algorithms that will probabilistically simulate the movement of people through the space. By adjusting global parameters (e.g. average work day, average time to commute to work, etc.) you should be able to simulate the movement of thousands of people through a space. This project might build upon the work done in traffic modeling (see MITSIM for work at MIT), modeling of behavior for virtual creatures (see Silas the Dog), and modeling of behavior for video games (see Maxis). Evaluation: Publishable paper on simulating the behavior of people that includes a general survey of prior work. Experience: Programming experience; interest in simulation modeling; interest in cognitive science. Transparent networking and video distribution using Java and C Level: UROP, MEng. Est. time: Hard to predict but probably around 1 semester. Start: Any. Description: The easiest way to provide a single video source to multiple computers is using video hardware such as video distribution amplifiers and quad splitters. This method is limiting, however. A better method would be to have code that can take video in from one computer and ship the video to another machine on a fast, local network. Develop code to do this in NT or find a reliable package that makes such video distribution possible. Java function calls should be provided to the low-level code that can be used by other application developers. Evaluation: It works (reliably!) without crashing despite hours of use and can be used by others easily. Experience: Programming experience; NT hacking; a knack for optimization; an interest in networked applications. Analyzing email usage and "constant connectivity" Est. time: 2 months. Level: Any. Start: Any. Description: How does "constant connectivity" provided by email and chat change the way that people are interacting with friends, family, and business partners? Use your observations to argue if and how communication will change in the future when people have higher-bandwidth (temporally and spatially) and wireless connectivity. Evaluation: A well-written, well-reasoned paper backed up with as many statistics on email usage as possible. Experience: Attention to detail. Context-aware faucet Level: Unknown. Research, software programming, and some construction. Est. time: 2 semesters. Start: Any. Description: The modern faucet is tremendously wasteful. Come up with a list of desirable properties for the faucet/sink. Using this list, design a new faucet/sink that has higher performance and that wastes less resources by using computational prediction of human action. Build a simulation for the prototype. Evaluation: The general design should be supported with a paper that argues why the device is an improvement over current devices. In addition, the prototype should be a reliable demonstration of the idea and ultimately something that could get incorporated into the House_n prototype. Experience: Programming experience; comfort with the goal of eventually controlling a small plumbing device by computer; a class in pattern recognition or computer vision; a class in AI. Implement "Eigenface" face detection for the prototype room Level: Any. Programming and possibly some research. Est. time: .25 semester. Start: Any. Description: Take existing code for face detection from images called "Eigenfaces" (because it uses eigenvectors) and incorporate the code into the House_n prototype space. This will require writing wrapper functions in Java for C code written by coders at Intel. Test the code in a "door" region of the room, so that the room can try to identify people as they enter the space from an image. Some modification to the code (or pre-processing of images) will most likely be necessary to get the algorithm to work in the House_n setting. Evaluation: A write-up describing the algorithm as implemented for House_n and testing it on a small dataset acquired in the space. The installation should run robustly and should provide hooks so that it can be used by other developers. Experience: C++ and Java software; image processing helpful. Tablet learning Level: Any. Design and some library research. Est. time: 3 months. Start: Any. Description: The invention of the blackboard transformed the way that information is taught. No longer did an instructor need to roam through students, looking at individual tablets (thereby paying individual attention to each student). Instead, the instructor could stand at the front of the room, isolated from the students, and present information. Can technology be used to re-introduce a more personal, tablet-like learning environment where learning takes place in a circular, intimate space? The goal of this project is to propose some ways of using technology to test this idea. Worthwhile would be interviewing a few teaching professionals and asking them to assess the new ideas. Evaluation: A series of well-reasoned ideas should be presented in a paper. Although the idea should be implementable with existing computer peripherals, the proposal may need to advocate the physical redesign of some components. If that is the case, design drawings and detailed descriptions of the interface to the design must be included. Background research on the impact of architectural space on learning should also be provided. Experience: Interest in learning and technology. Child monitoring Level: Any. Est. time: Depends on scope. Start: Anytime. Description: Using sensing technology available in the House_n prototype space, consider how it might be used to monitor the activity of children to provide some information to parents. Evaluation: Depends on scope. Experience: Depends on scope. The virtual meal Level: Any. Video observation study with some equipment setup. Est. time: 2 months. Start: Anytime Description: The purpose of this project is to investigate how remotely linked environments will change interaction in homes of the future when infinite bandwidth is a reality. Simulate infinite bandwidth between the prototype space and the satellite space (located in a nearby office) by running some video cables. Create a scenario where people are eating dinner in the prototype space and people are conversing with them in the satellite space. Evaluate the problems you encounter. Propose some interface ideas to improve the interaction (e.g. by overlaying information on the views, etc.). As much as possible, construct real scenarios where people on both sides of the space are doing real activities and closely observe the results. Evaluation: Report on the problems experienced with suggestions on how to fix them. Interviews with participants. Experience: Interest in virtual communication; communities. Collaboration from a distance Toolkit for developing ambient displays Apartment messaging system Level: Unknown. Est. time: Unknown. Start: Unknown. Description: Create a system that uses the PDA and camear to allow people to take pictures of their environments and label them for others who maybe seeking information about that spot later. In particular, this is for people purchasing or renting property and a way for them to get space-based information. User can "augment" camera images. Picturess do the work. (In ultimate version camera might know the 3D location). Take a picture and pops up picture: what do I know about this? See page in notes folder. Evaluation: Unknown. Experience: Unknown. Background subtraction and object tracking in doors with natural light Level: UROP. Est. time: 1-2 semesters. Start: Any. Description: Nearly all computer vision systems run indoors make unrealistic assumptions about light. Assume that you have a room with lots of windows looking outdoors. Starting with existing background subtraction algorithms, modify them so they can handle natural lighting situations (e.g. clouds passing overhead that instantaneously changes the light in the room coming from the windows). Extend your work to use existing tracking code for the House_n environment. Evaluation: A paper describing your algorithm and testing it thoroughly. Experience: Java/C programming; some background in image processing/computer vision helpful but probably not required. Ultrasound triangulation pointing device Level: MEng electrical engineer or higher. Est. time: 1 year. Start: Anytime. Description: Design a remote pointer control device using ultrasound that will make it possible to create a "touch screen" on any flat surface in an environment by installing multiple copies of this device. A pen would hold a small transmitter. On each large flat surface, small receivers would be attached. Hardware would perform triangulation to determine the position of the pen on the wall. Products exist that do this (e.g. Mimio and EBeam) but they all have problems when you try to apply them to the home of the future because they were designed with different applications in mind. Evaluation: It would need to work and we would need to be able to create multiple copies of the device. Experience: This project will require significant experience developing hardware that does fast signal processing and significant comfort writing low-level software drivers. Might best be approached in a team. Robust, continuous people tracker Level: MEng+. Research and software programming. Est. time: 6+ months. Start: Anytime. Description: Develop a robust, continuous people tracker from video input that can run for days without crashing. Evaluation: Evaluate the algorithm on test data and publish results. The system should be easy to use by other application developers for the prototype space. Experience: Software programming; course in image processing. Robust Java/WinNT system architecture and administration Level: UROP. Est. time: Unknown. Start: Anytime. Description: Help to design a rock-solid Java/WinNT system architecture to be used by a an interactive room prototype that can run for days without crashing. Should provide RPC-like message passing between individual computers and processes and handle expected problems like power failure while still rebooting and running. Evaluation: Runs for days without crashing and is easy to plug other code into. Experience: Expert WinNT programming. Class input Level: UROP, MEng. Programming. Est. time: 1 semester. Start: Anytime. Description: Using off-the-shelf WinCE PDAs, create software so that up to 20 PDAs can be hooked up to a computer via serial port or wireless links and a lecturer can ask a multiple choice question and get input from each of the students. The device would be used for an interactive teaching task. Time permitting, use the device to in a user interface application to demonstrate a new interactive teaching idea. Evaluation: Construct a functional prototype using at least two PDAs. The software should be well documented so someone unfamiliar with your code can use the code with new questions. Experience: WinCE programming, Java/C programming. Detecting dance (and other) activity Level: MEng. research. Est. time: MEng. Start: Anytime Description: The goal of this project is to use sensors in the House_n ubiquitous computing living room to detect specific events and one in particular: if people are dancing or not. The detection of this event could be used to change the way the space responds using ubiquitous audio and video. The project requires identifying contextual cues that indicate dancing using as a toolbox a set of visual primitives and an activity map of behavior in the home. One possibility is to us a supervised learning algorithm, which would require developing a natural interface by which occupants could interact with the system to provide training information. This project will require significant experimentation. Evaluation: The algorithm will need to be tested using a small dataset of new examples. The use of the algorithm and the training interface (if there is one) should be demonstrated. Experience: Some background in pattern recognition (e.g. a PR course) or designing sensing algorithms. Proficient in Java and C programming. Research Y2K predictions
	Level: UROP.
	Est. time: Partial semester.
	Start: Anytime.
	Description: Compile a list of "future" predictions and analyze why or why not the predictions hit the mark.
	Evaluation: Report.
	Experience: Interest in library research.

Your project description here?
	Level: Unknown.
	Est. time: Unknown.
	Start: Unknown.
	Description: What is the analogy of Google algorithm for place-based web (distributed, geographic web)?.
	Evaluation: Unknown.
	Experience: Unknown.

Your project description here?
	Level: Unknown.
	Est. time: Unknown.
	Start: Unknown.
	Description: Unknown.
	Evaluation: Unknown.
	Experience: Unknown.