HANDWRITING RECOGNITION: Linderman and his colleagues developed a test glove for their "digital hand" experiments and used a Wacom Intuos3 digital tablet and pen to help them reconstruct the actual movement of the pen. Image: Courtesy of Norconnect, Inc.
A team of researchers is looking to handwriting rather than the ever-shrinking keyboards (that frustrate typists) in a quest to improve the way we input information to computers and cell phones. They say the key (especially for the less nimble among us) is "digital hand" technology, a glove stocked with embedded sensors that translates into digital text the electrical impulses generated by hand muscles during the act of writing—on any surface, using any pen, pencil or even a finger.
Digital pens that can read and store a person's handwriting as digital text have been around for years. But Michael Linderman, president and chief technology officer for Norconnect, Inc., the Ogdensburg, N.Y., company developing the Digital Hand miniglove, wants to provide an alternative that is not reliant on special writing devices or paper.
In the first phase of his research, Linderman, along with Mikhail Lebedev, a senior research scientist in Duke University's Department of Neurobiology (with expertise in modeling and statistical analysis), and Joseph Erlichman, a neuroscientist and professor of biology at St. Lawrence University in Canton, N.Y., (whose specialization is hand muscle anatomy, physiology and electromyography, or EMG), conducted a series of experiments to better understand how the hand's muscles work during the act of writing. As with any bodily movement, muscle tissue contracts and emits electrical signals that are sent to the brain when pen is put to paper.
To read these signals, the researchers created a prototype glove with electrodes positioned over different muscles in the hand and connected to a computer. As the hand moved, signals were recorded via EMG from the different muscles involved in the process. When researchers examined the results of these tests, they found patterns unique to individual characters that the hand had drawn.
Linderman's goal was to use the data to predict the trajectory of the pen while the hand was writing, says Andrew Fuglevand, an associate professor of physiology at the University of Arizona College of Medicine–Phoenix with whom Linderman consulted during his research. (Linderman also consulted Carlo De Luca, director of Boston University College of Engineering's NeuroMuscular Research Center.) "From there," Fuglevand says, "he would be able to figure out which character the pen was drawing."
Linderman and his colleagues opted to use skin-surface electrodes for their tests. Although an EMG measurement is more accurate when electrodes are inserted directly into the muscles being studied, any handwriting recognition device they ultimately created would have to read signals unobtrusively. Linderman was able to identify different patterns in the signals picked up by the electrodes and associate those patterns with the different characters that had been written. "All he was trying to do was look for a particular signature in these patterns," Fuglevand says, "and he was able to extract information in an impressive way."
Fuglevand's own research relies on EMG to study patterns of muscle activity associated with different movements, with the goal of better understanding how to reanimate limbs in paralyzed individuals. When someone has a spinal cord injury, he says, the muscle is still viable, but the signal cannot get from the brain to the muscle. Linderman's work can essentially be reverse engineered to help with Fuglevand's research. "This will have influence on the development of so-called neural prosthetic devices," Fuglevand says. Linderman "is recording electrical activity from many neurons and from that predicting what a person is doing in terms of moving their limb or tracing out a letter."
"At this point, we've proved the concept and are going for phase two to develop the complete prototype," Linderman says, adding that he paid for phase one with a $100,000 National Science Foundation (NSF) Small Business Innovation Research (SBIR) program grant he received in July 2007. Now he is looking to the NSF for a second round of funding to the tune of $500,000 that would help him build a more advanced prototype within three years. Whereas the current prototype is physically connected to a computer, the next generation would transfer data wirelessly via Bluetooth to a computer or mobile phone. Ansen Corporation in Ogdensburg, N.Y., will help the researchers manufacture and distribute the technology when it is ready for broader use.
Digital Hand will be an alternative to the "smart pen" and touch-screen finger writing technologies. Livescribe's Pulse and Leapfrog Enterprise's FLY Fusion are high-tech pens that rely on sensors and special paper with embedded dots to capture and digitize a user's handwriting. Meanwhile, mobile phone makers including Motorola, Pantech Wireless, LG Electronics and HTC Corporation sell handsets that allow the user to write directly on the phone's screen. But these units cost a few hundred dollars more than the average phone and are not all available in the U.S. Linderman says these technologies don't get it quite right, because they replace the keyboard with technology specific to the individual device. "The user should have the freedom to choose any pen or paper to write with," he says.
Linderman anticipates the Digital Hand glove will be priced competitively with digital pens. "Our market research indicate[s] people will spend about $100 on it," he says.
Another project Norconnect is pursuing with the help of the Frederic Remington Art Museum in Ogdensburg is an application of the digital hand concept to drawing and other artistic endeavors, in addition to handwriting. An EMG-based device will provide artists with a new tool for developing their creativity, Linderman says, by capturing and interpreting signals from various arm muscles to create a drawing. "Our intent is to study the two-dimensional drawing first," he adds, "and then progress to a three-dimensional artistic hand and arm motions in the next project."
In the end, Linderman is also concerned with how keyboards have changed the way people communicate, making cute abbreviations like "lol" ("laugh out loud," for you traditionalists) and "omg" ("oh my god") preferable to more nuanced language. "People have begun to change their language to accommodate technology," he notes, when it should be the other way around.