EECS 373 Fall 2011 Sponsored Project Opportunities Prabal Dutta, Matt Smith, Lohit Yerva, Darren Ashton, and Pat Pannuto Background: There are lots of interesting embedded systems out there that researchers could use but either don't have the skills or time to actually design. However, some of you many have the time and skills, especially after 373, to create such things. We'd like to encourage that by identifying some projects opportunities. Most of these won't be as cool as building a small Segwey or a quadcopter. On the other hand, they are quite useful for addressing a range of research problems. Picking one of these projects will ensure that you will get a lot of attention from the 373 instructional staff as well as a nearly limitless budget (at least by 373 standards) to carry out your ideas. All of these projects can easily lead to publications in top tier research conferences, so if you're interested in graduate school, this would be an excellent way to get some visibility. 1. Talking Book. Modify an existing PCB to add power supplies and other electronics. Write software drivers to implement wireless communications (over a magnetically-coupled interface) between devices. Verify and characterize the communications interface (power, throughput, etc.). This work, if successful, will be designed into next-generation talking books and disseminated widely (hopefully in the millions) targeted at helping illiterate populations in the developing work. See the following paper for a background on the project: http://www.eecs.umich.edu/~prabal/pubs/papers/foo10book.pdf 2. @fingertips. Design and implement an iPhone augmentation device that allows people with severe visual impairments to use the phone. This project's industrial design has already been done by students in SI and the Ross Business School. Your team's role would be to build a prototype including hardware and software. This would involve both embedded systems and mobile phone development. See the following link for more information: http://www.youtube.com/watch?v=R8X2d0WnwgA 3. Smart badge. Design and implement smart badges that can be worn by people and collect human inter-contact data. A functioning system will capture any interactions that last longer than about 10 seconds between two closely-located (roughly 0-2 meters separation) people. The system should be small (less than 2" x 2"), be able to run off a coin cell battery (250 mA-hr) for at least a few days, and cost no more than $25 in modest volume. It should include at least a few MB of NOR flash memory. A successful project could lead to a large-scale (1000-2000 node) deployment in the state of Utah in elementary, middle, and high schools. The following paper provides some background on what we're trying to do but the goal is a smaller, less expensive, and more useful badge: http://sing.stanford.edu/pubs/hotemnets10-flu.pdf 4. Power Line Communications. Integrate a power line communications (PLC) module with a SmartFusion system. Design a PCB that includes a wall-plug interface. Write the drivers for the PLC module in C and demonstrate communications over building wiring. This project, if succesful, could have industry impact as people are just now starting to consider using PLC in wireless sensor networks. See the following paper for some results from our preliminary work: http://www.eecs.umich.edu/~prabal/pubs/papers/pannuto11plc.pdf 5. Microcontroller and radio for a cubic inch wireless AC power meter. Last fall, one of the teams designed and implemented a wireless power meter that eventually became just a cubic inch form factor. This term, we'd like to convert the processor and radio of this system to use an ARM processor. This involves designing a small PCB, less than 0.75" x 0.75", and getting the software drivers for the processor, radio, and power meter working. All of the other electronics have been designed already. You have two options for the radio: either 802.15.4 or 802.11. The 802.11 radio would be preferred as it's most different from what we have already. If you build this, lots of people will use it both at Michigan and elsewhere. Email me about this project for more details. 6. HiJack++. Port the HiJack system (currently running on a TI MSP430 processor and iPhone handset) to use the EnergyMicro TinyGecko ARM Cortex family. Ideally, port the HiJack software onto Android or Windows Mobile (of course we'll provide the phones). See the following page for some more information: http://www.eecs.umich.edu/~prabal/projects/hijack 7. AutoWitness Tag Sensors. Integrate an ARM microcontroller, GPRS radio, flash memory, accelerometer/gyroscope, vibration sensors, and battery into the smallest possible form factor and make the system draw practically zero power when sitting still. This design, if successful, will be used in Memphis, TN to catch thieves in the act. Much of the back-end software already exists so the main challenge here is to get to a viable platform. See the following paper for some additional details: http://www.eecs.umich.edu/~prabal/pubs/papers/guha10autotrack.pdf