The Acceleration Sensing Glove (virtual keyboard) was designed to demonstrate that accelerometers can be used to decipher and translate hand gestures into computer interepreted symbols.  The glove and software were designed by Seth Hollar, John Perng, and Brian Fisher in Professor Kris Pister's lab at the Univeristy of California, Berkeley.


Glove links:

How Does It Work?

An Analog Devices 2 axis ADXL 202 accelerometer is placed on top of each finger on the glove.  Additionally, a sixth accelerometer is placed on the back of the hand.  The analog signals from the accelerometers are digitized by an Atmel AVR microcontroller, which in turn sends the data to the computer via the serial port.  The hardware supports wireless transmission of the data over RF but currently that capability has not been demonstrated.

The raw signals from the accelerometers can be translated into rotations the hand makes with respect to gravity (see figure on left).  These rotation angles are then used to describe a particular finger orientation.   The computer tries to find the hand gesture that possesses similar information for each rotation angle. That way the computer can interpret a particular hand gesture as a unique symbol.

It should be noted that currently we have been looking into recognizing static hand gestures (still hand gestures).  Examples of static hand gestures might be the peace sign, thumbs up, a flat hand, and most of the american sign language alphabet.

What Can We Do With It?

Based on the description above we have developed an application that recognizes hand gestures.  The application can recognize almost any hand gesture you train it with.

Additionally, we have developed an application that allows you to use the glove as a mouse pointing device for a Windows 95 system.  By tilting your hand to and fro you can move the pointer in the same fashion as any other mouse.  Button clicks are handled by pointing certian fingers down.


 

What Now?

The Acceleration Sensing Glove is a large-scale device used to model the functionality of Smart Dust on fingers.  Smart Dust is a project whose 3-year goal is to integrate communications, intelligence, power, and sensors into a package that is no larger than 1 mm3 (See Figure Below).  Integrating a single chip wireless solution with a MEMS accelerometer would yield an autonomous device small enough to apply to the fingernails.  Because of their small size and weight, these Smart Dust devices would be less noticeable than one?s own eyeglasses, providing no more discomfort than fingernail polish.  People would have instantaneous input access to the digital world at all times, facilitating a paradigm shift in human-computer interaction.

Some potential applications for the  Acceleration Sensing Glove are: American Sign Language Alphabet interpreter, a wireless wearable mouse pointing device, a wireless wearable keyboard, hand motion and gesture recognition tools, virtual musical instruments, computer sporting games, and work training in a simulated environment.
 



Conference Paper for ISWC International Symposium on Wearable Computers, San Francisco, October 18-19th.

Questions?  Comments?  Please contact Seth Hollar.
 

This site is a mirror of http://www-bsac.eecs.berkeley.edu/archive/users/hollar-seth/fingeracc/fingeracc.html