Assignment #2: Research Safari

These are the five most relevant research projects I found at the EECS 2000 Conference.  Since my background is in hardware, I spent my time at the poster sessions in Soda catching up on the activity going on in the CS division.

1. Human-computer interfaces
   I think this is absolutely fundamental to the vision of environment-integrated computing assisting or enhancing human activities. Without the proper human interfaces, we will never get 'outside the box' in interactive applications. There are a bunch of highly relevant topics in this category, including speech, handwriting, and gesture recognition.
2. MEMS sensors
   To support 'smart spaces', we must have devices that can efficiently detect the presence and location of objects. Extremely cheap, small, energy-efficient devices are needed to perform this location service. MEMS technology enables all three qualities, which are essential to provide thousands of sensors in a building, or to make them highly mobile and disposable.
3. Wireless communications
   If any of the deeply networked devices we conjure up need to be mobile, they will need wireless communication. As the number of communicating devices gets higher, and the data rates required by applications get more demanding, both the RF hardware and communication algorithms and protocols must be designed to efficiently and securely transmit the data devices need.
4. Lightweight databases and/or property-based document organization
   There is no umbrella research group to cover these technologies, but I found them both somewhat related and relevant. The task of accessing and querying the mind-boggling amount of data collected in a smart environment will be very important. This could be on-the-fly data processing for extracting content from a meeting in progress, or for cataloging information once the capture has ended. I thought that either a highly efficient but classic database access protocol or a new method for accessing documents based on content would help this issue.
5. CAD Tools
   Yes, this seems a bit random. But the level of integration of modern chips has rendered design tools almost incapable of producing next-generation devices without huge amounts of effort. If we are going to place thousands of devices in the environment, we need new tools that can rapidly and easily design devices of this nature. Not to mention, if we do try to make every object in the environment electronically intelligent, we need to make it possible for smaller companies to design them. It shouldn't be necessary for a company to use multi-million dollar design suites to get a chip out to a foundry.