APPLICATIONS

By Nancy Gores – Trace R&D Center, Univ. of WI, Madison -- The dynamic environment of the Access Grid (AG) could offer unprecedented higher education and collaboration opportunities to a population that is often overlooked. This group is people with disabilities—54 million strong in the United States, and about 750 million people worldwide. The ensemble of resources on the AG that supports human interaction—multimedia display, presentation and interaction environments, interfaces to visualization environments—are powerful communication tools. But what if you've made a career studying materials at the nanoscale and you suddenly lose your vision? What happens if you're a successful chemical engineer who experiences a hearing loss? What if you possess the intellectual depth to become a successful scientist, but are born with a vision or hearing impairment that severely limits your access to the world of knowledge? How can you participate in collaborative scientific environments given these constraints? The Trace Center, University of Wisconsin-Madison, an EOT-PACI partner, has been doing research and development in the area of disability access to technology for 30 years. Through its work within EOT-PACI and the Universal Design and Disability Access project, Trace is making collaborative environments accessible to people with different abilities. This work, which requires researchers to develop flexible features to accommodate different users, will also make the AG more usable to people who are not disabled. Modality Translation At a Birds of a Feather (BOF) session at SC2001 in Denver, participants discussed services that allow blind people to hear descriptions of visual information, and allow the deaf and hearing impaired to access real-time text transcripts of AG events. The session was a live captioned example of an accessible service using an experimental AG-based speech-to-text translation service. The remote audience gathered at AG nodes at NCSA in Urbana, IL and at the Alliance's ACCESS center in Arlington, VA. Speaking in Denver, UW-Madison Trace Center Director Dr. Gregg Vanderheiden presented the audience with information about the use of Grid-based modality-translation services to provide access for individuals in constrained environments and for people with disabilities. The text captioning experiment was set up to record live audio from Denver and transmit it to the Trace Center in Madison, where a student provided speech-to-text translation by re-voicing into off-the-shelf speech recognition software. The text was displayed in large letters across a wall at the Denver location. This experiment was designed to explore network related issues with a Web-based prototype developed at Trace, and to demonstrate potential technical solutions for modality translation on demand. A network difficulty caused the audio to break up badly, and as a result, participants could not hear the presentation, nor could the sign language interpreters in Virginia translate it. The problems demonstrated the necessity of providing full text transcriptions of Access Grid events. With a professional text transcriber providing high-quality text transcription, people with hearing impairments could participate in any AG sessions without the need for sign language interpreters. In the event of audio difficulties the text "captioning" would be available on demand for all users as well. At the end of the session questions from the audience were answered by Vanderheiden, Al Gilman and Gottfried Zimmermann. Gilman and Zimmermann are researchers working within EOT-PACI on the Universal Design and Disability Access project at the Trace Center. Due to the audio difficulties, questions from the remote sites were asked via typing into the chat window of the speech-to-text clients. From a technical perspective the experiment with the text transcription prototype was successful. Each of the three connected nodes ran at least one client (applet) of the Web-based prototypical system. The translator's text stream flowed continually on the speech-to-text windows. Visions for the Future The BOF participants at remote sites used a dedicated laptop computer to type questions and comments to the rest of the group. Plans are to improve this system by enabling all participants to use their own computers and hand-held devices to participate in collaborative sessions. The AG's plans for a "docking" functionality could play a key role in meeting this requirement. Docking would allow participants to use their laptop computers and handheld devices to wirelessly dock into the collaborative environment. Also, sometime in the future, a sign language recognition system could translate the signs to text in real-time. However, research in the area of sign language recognition is just beginning and it will be some time before low-cost systems are available to deliver this service. An accessible Access Grid, facilitated by modality translation services such as the speech-to-text service, would not only benefit people with disabilities. Every user would benefit in situations where communication capabilities are constrained. For example, a scientist could participate in a collaborative session while driving and receive verbal descriptions of a diagram discussed in the session. Moreover, an accurate text transcript of the meeting would allow for indexing of recorded audio tracks. As part of the NSF-funded PACI program and its Education, Outreach, and Training efforts, the Trace Center is working on a concept called "Modality Translation Services on Demand." This concept stretches beyond the Access Grid as an advanced collaborative environment. The concept could facilitate a grid where everything is accessible to everyone, even dynamic information such as real-time collaborative experiments and simulations. Story used courtesy of NCSA's Access