Like many people with Duchenne muscular dystrophy (DMD) and other neuromuscular disabilities, I've tried many assistive technology devices for accessing the computer. Some didn't work for me and others only worked for a short time until the progress of my disability outstripped them. But with the recent improvements in computer assistive technology and a little persistence, I've been able to find a combination of solutions that's worked well.
I began using assistive technology devices for the computer in 1984, when I was in college. As my DMD affected my hands, I started to lose the ability to use the keyboard. I knew that the computer would be an important tool for me in school and for employment as a software engineer. And I knew I'd need AT to allow me to use the computer effectively.
|Scott Bennett and his pug-beagle-mix dog, Peanut, have tested the limits of voice recognition software. Bennett has his home computer set up to use with a pointing stick and a wired microphone.|
I first experienced computer assistive technology at the University of Massachusetts at Boston, when I volunteered at the Adaptive Computer Lab, which provided resources to students with disablitities.
At that time there were very few devices for the computer around and what was available was in its early stages of development. The Adaptive Computer Lab had screen magnifiers and speech synthesizers for visually impaired students and was experimenting with miniature keyboards and voice recognition technology for students with mobility impairments.
I started using a miniature keyboard that required very little range of movement and allowed me to enter keyboard commands without much effort. This 4-by-7-inch keyboard had a plastic membrane keypad with keys printed on the membrane, very similar to the USB minikeyboard currently made by Tash (www.tashinc.com).
The miniature keyboard was an effective means of using the computer with the limited mobility I had in my hands and arms, but it did have some limitations. This early version of the membrane minikeyboard would occasionally misinterpret or fail to recognize a selected key, didn't have mouse control and connected to the computer using the serial port. (When I started using computers the mouse was just beginning to be used and USB hadn't been invented yet.)
Today, the USB minikeyboard and many other alternative keyboards provide mouse control and can improve access to the computer for people with limited hand and arm mobility.
I could've used the minikeyboard for several years, but I knew that as my DMD progressed I'd eventually need to find an alternative to it. So I began looking at voice recognition software to access the computer, even though it was a relatively new technology at the time.
Voice recognition technology, or more accurately speech recognition technology, requires you to speak into a microphone connected to your computer and breaks down your spoken words into phonetic sounds that the computer in turn interprets as words; the computer then translates the words into text, keystrokes or commands.
I got my first voice recognition software product in 1986, made by Kurzweil AI. This DOS-based software was very primitive compared to what's available today, having a vocabulary of only 1,000 words and recognition accuracy of only 70 percent. (OK, so I was using computers before Windows software and enhanced graphics!) Although this software allowed you to speak to your computer, the limited vocabulary meant you could only send basic commands.
You had to speak the military alphabet to type letters on the computer: "alpha" for the letter A, "beta" for the letter B and so on. You had voice commands for other keys, like "enter," "spacebar" and "tabkey," and for combination keystrokes like "control-charlie" for Ctrl-C. With voice commands for all the possible keystrokes there was little room left for voice commands to open and close programs, use menu items or use a vocabulary when dictating.
This early voice recognition software was very slow compared to typing and made many mistakes interpreting spoken words. Writing a paper took a lot longer since you had to correct the misinterpreted words. But I didn't give up on voice recognition technology because I knew the software would advance to include larger vocabularies and better recognition accuracy.
The most significant problems with the early voice recognition software were the limited vocabulary, limited accuracy and the inability to control the mouse. But soon after I began my career as a software engineer, new versions appeared that addressed those problems. With the introduction of the Internet and advances in software engineering and with the progression of my DMD, these improvements would be vital for my continuing to work.
These improvements didn't come without setbacks, however; technological problems and solutions may be technical or organizational - or both.
When Kurzweil AI was bought by Lernout & Hauspie in 1997, the company released a new version of the voice recognition software that no longer provided mouse control by voice, an important option for me. So I began creating my own mouse application. But before I completed it a dramatic improvement in voice recognition technology emerged.
After buying Kurzweil, Lernout & Hauspie also bought Dragon Systems, which had its own voice recognition software called DragonDictate. ScanSoft (now called Nuance [www.nuance.com]) bought out Lernout & Hauspie and in 2000 decided to merge the best aspects of Kurzweil with the best aspects of DragonDictate, to create a new version of Dragon Naturally-Speaking.
The new product improved the recognition accuracy to 95 percent, included a 300,000-word vocabulary, reintroduced mouse control by voice and introduced continuous speech capability.
Continuous speech technology, or natural language processing, allows you to speak to the computer without pausing between words, as if you were talking to another person. With continuous speech, Dragon NaturallySpeaking gives you the capability of dictating e-mails, documents and text much more quickly than with discrete speech, which requires you to pause between words. But Dragon NaturallySpeaking also gives you additional modes for entering commands or individual letters when you don't need continuous speech capabilities.
I've been using Dragon Naturally-Speaking for over six years now and have found it to be an excellent product. But despite all the impressive claims and capabilities, voice recognition software isn't perfect and often makes mistakes.
For example, every time I say "keyboard" the computer writes "keyword." It's a significant editing chore to correct every instance when the voice recognition software makes that mistake.
The amount of ambient noise can also cause the software to make mistakes, and you may have to do a microphone check to let the computer know that the background noise level has changed.
There are times when I'm at home talking to my computer and my dog will start barking at something he sees outside; the computer will start typing the letter R. So, according to Dragon NaturallySpeaking, a dog barking is R, R, not woof, woof.
Although these problems can become annoying and frustrating, the recognition accuracy is much better than it was when I first started using voice recognition software. And the improvements in vocabulary size and continuous speech capabilities have made using the software much more effective and efficient.
While Dragon NaturallySpeaking is my primary means of accessing the computer, I also use other AT products along with it. For many years I've been using the Microsoft Accessibility Tools for the not very rare occasions when Dragon Naturally-Speaking, like many Windows-based software products, crashes or starts running slow.
The Microsoft Accessibility Tools are now standard Windows components that provide easier access to the computer with some useful utilities. I frequently use Sticky Keys and Mouse Keys.
Sticky Keys gives you the ability to type combination keyboard commands like Ctrl-Alt-Delete one key at a time by remembering the individual keys that are pressed. And Mouse Keys allows you to use the keyboard's numeric keypad as a mouse.
I use these tools by pressing the keys with a pointing stick attached to my wheelchair. The pointing stick is often a topic of conversation when people ask what it's used for, and before telling them the real reason, I usually think of something creative to say like "crowd control."
Other useful Microsoft Accessibility Tools utilities can be activated by selecting the Accessibility Options icon from the control panel of the Windows XP operating system.
The most recent product I've added to my computer is a wireless microphone that allows me to speak to the computer without having to be attached to it.
There are a couple of wireless microphones available that can be used with Dragon NaturallySpeaking. I chose the Plantronics (www.plantronics.com) CS50 USB wireless microphone because of the cost and ease of use. Its single drawback is that the batteries will last only eight hours before needing a charge.
I considered using the wireless microphone made by Shure (www.shure.com) because the batteries lasted 12 hours, but it was more expensive and seemed too cumbersome.
Using these computer assistive technologies together, I have complete access to the computer, any computer application and the Internet. But finding the right assistive technologies for me has meant spending a lot of time, effort and perseverance as well as anticipating my future needs to ensure that I have the access I need.
The computer assistive technologies I've used over the years have provided me with many opportunities for school, for work and at home. And now, at age 42, I couldn't maintain my career as a software engineer without the help of these technologies.
Scott R. Bennett is a software systems engineer in Taunton, Mass.
Adaptive Computer Products
Ability Hub: Assistive Technology Solutions
RJ Cooper & Associates
IBM Accessibility Center
World Wide Web Consortium (W3C)
Web Accessibility Initiative
Section 508 Amendment to the Rehabilitation Act
Dragon NaturallySpeaking by Nuance
IBM ViaVoice by Nuance
QPointer by Commodio
Speak to Me by RJ Cooper & Associates