IT Technologies, and the kind of things you can do!

     My professorship at the KTH is in IT product development, which is an area that I enormously enjoy working in.  I received a Ph.D. in Bioengineering from the University of Utah in 1983, and joined Hewlett Packard Labs in 1984 after a year as a visiting faculty member in the University of Utah Electrical Engineering department. Working at a company like Hewlett Packard was a great way to not only learn from some of the best engineers in the world, but it also gave me a chance to see how technology goes from being an idea to real technology and finally developed into a product.

      While I was at Hewlett Packard Laboratories, I was involved in a large number of different technology research areas.  My most recent work areas have been in new personal devices that exploit both the Internet, and context awareness.  One nice thing about having a Bioengineering background is that the role of sensors in human systems form a natural input to future service based computing architectures. The SmartBadge architecture series has been a central platform for a number of projects and experiments in context aware computing, mobile media devices and low power systems.  A number of interesting platforms and systems have spun off from this work, including new kinds of user interfaces based on sensors, security devices for identity management and media enhanced communication.  The last version of the SmartBadge, shown in the photo on the far left,  was especially fun as it supported multiple radios for communication, ran Linux, had USB host ability, and had a number of hooks to make it easy to measure the power consumption of the various subsections, such as the radios, memory system or processor core.  The SmartBadge shown is being used in the original prototype of a very interesting project that was called the DJammer interactive MP3 architacture.  The DJammer allow people to interact with music streams by performing time shifting, or 'scratching' in the same way a professional DJ does.  Because it is wireless, it allows the resulting sound to be shared with others.  The picture next to the SmartBadge shows a professional DJ (DJ Gavin O'Conner) using the final version of the DJammer.  He is actually scratching on a sound track by moving his hands in mid air in a very similar way that he would move his hands on a turntable.  The sensors used to detect his hand motion are contained in the metal handpiece  which serves as the entire user interface, and is an example of a really beautiful mechanical design. Projects like the DJammer are great to do as they show examples of the role of sensors and sensor support infrastructure in future consumer devices.  They also show how media can really be used to communicate, inform and cause delight.  I think the term "delight" is a good measure of consumer product success, and should be included along with volume and profit.

     Other technology areas I've worked in have been in new imaging devices.  One of the coolest systems that has come out of HP labs was the Capshare handheld document scanner shown over on the right.  The name Capshare is a contraction of the words Capture and Share, but the internal code names for the device were "Swipe" and later on "Zorro", and were much more evocative of how it was used than the product name.  The basic idea was a standalone handheld device to scan in any document or other artifact on a flat surface at a resolution of 300 dots per inch using no moving parts and completely natural hand and arm motion.  To use no moving parts was a challenge in itself, but using natural hand and arm motion meant that the images could be very spatially distorted due to highly curved motion trajectories.  It also had to work successfully at a velocity of up to 1 mile per hour over the surface being scanned, and had to fit in your shirt pocket.  I worked with some of the best engineers and computer scientists in HP on this project.  Capshares are now collector's items and you can only find them on ebay, but much of the basic technology that came out of the Swipe project can be found in other devices today.

     Communication infrastructure technologies are also areas I've done research in.  One of the visions shared by a number of researchers in HP labs has been the use of video as a communication modality, especially in emerging wireless LAN systems.  In early work at HP labs we looked at how 3D graphics and video could be composited and sent over a wireless channel.  This work gave rise to the MINT Wireless Mobile Internet Router, which was work done with the KTH.  A lot of research was done in multiprocessor systems to support algorithms for 3D graphics rendering and solid modeling.  As is still the case today, one can never really have enough performance in graphics systems.  Some of my current research in communication infrastructure is centered around how to aggregate multiple devices into forms that applications can discover and use as if the aggregation were a single functional device.

     My vita gives details about some of the systems that have been designed in these areas, plus pointers to publications and patents that have come out of this work.  My favorite patent is the one titled Method and Device for Tracking Relative Movement by Correlating Signals From an Array of Photoelements.  This is an all optical motion sensor that was intended for use as the 2D surface navigator in the Swipe hand held image scanner.  The technology in it gave rise to all the sensors used in today's optical computer mice.

     I'm always ready to talk about technology, research ideas, or what life is like in a large industrial research lab.  Come by and visit.  My office is in Electrum 1, 3rd floor, elevator C.  Go into Wireless@KTH and then to room 6345, or send me email at msmith #@# kth.se.