Telecommunications Advances Helping to Make Equal Access a RealityDecember 2020
Topics: Communications Technologies, Communication Services, Communications Switching
Remember that time when you had no cell service and desperately needed to connect with family and just couldn’t? What if that were your experience every moment of your life?
Nearly 48 million Americans have some degree of hearing loss. Many individuals have been deaf and/or Hard of Hearing (DHOH) all their lives. For those who can hear, it’s impossible to imagine the frustrating roadblocks one faces in the non-hearing world. At some point, many of us will cross the divide and be faced with the reality of hearing loss.
The Federal Communications Commission (FCC) has provided improved access to telecommunications through the Telecommunications Relay Service (TRS) program since the 1990s. In 2015, the FCC established a relationship with MITRE to research ways to improve telecommunications access and outcomes for this population.
MITRE, working in close partnership with the FCC, has been crucial to recent improvements in the agency’s TRS initiative. The result? Increased accessibility, expanded services, and improved quality.
Several new service options based on our research have and will continue to help broaden the market, improving the efficacy and efficiency of the program and further enhancing accessibility.
“Our work with the FCC and the TRS initiative spans several areas, and our efforts to better understand and help address communication challenges have resulted in real impact,” says Reeta Singh, the MITRE project lead.
Beyond cost savings and increased usage, the FCC made two data-driven policy decisions over the past year that will expand the provider market and will significantly improve the ability of the DHOH community to connect using technology.
“These were a direct outcome of MITRE’s innovative development, research, and testing,” Singh says.
A Focus on the Human Side of Technology
When MITRE’s collaboration with the FCC on the TRS initiative began, TRS consisted of three services that relied on “relayed” communications. Users with hearing or speech disabilities could communicate via technologies that relied on human communication assistants (American Sign Language, or ASL, interpreters) in Video Relay Services (VRS); revoicing in Internet Protocol Captioned Telephone Services (IP CTS); and text to speech in IP Relay.
These services, while largely acceptable to the users, were often labor-intensive for both users and providers. Moreover, challenges and opportunities existed: privacy could be improved, performance of the technology and/or interpreters was inconsistent, and previous research was proprietary to the providers. That meant it wasn’t widely available to the users or the FCC.
One of the first areas MITRE and the FCC explored was automated speech recognition (ASR) systems. Early ASR engines offered potential cost savings but required further development and intense scrutiny to ensure parity in service compared to existing captioning offerings.
To address these issues, MITRE engineers developed a multi-ASR captioning prototype engine called ACE Quill. We then used our prototype to test how well existing ASR engines performed, as determined by Hard of Hearing users.
“By gathering data on factors that help and hinder the use of captioned phones, we learned firsthand about many different communication challenges people with hearing loss face using a phone,” says Mark Pfaff, who leads MITRE’s usability team.
“We worked with academia, industry, and especially with users of hearing assistive technologies. We wanted to help ensure our usability research and experimentation wasn’t just about technology—it’s mainly about the people who use it.”
With the ACE Quill prototype, the MITRE team could quickly assess performance baselines of emerging ASR engines and determine a method for comparison to existing solutions.
The results of our research and testing found that removing the “revoiced” component of traditional IP CTS improved lag time of the captions by at least 50%, without a significant corresponding drop in accuracy. In fact, we discovered that caption presentation, accuracy, and delay played a significant role in users’ willingness to adopt ASR technologies.
We also learned that for a portion of the Hard of Hearing population, audio amplification was more effective than the added task of reading captions during a conversation.
“Knowing that our development of ACE Quill has led to real-world improvements and stronger communication capabilities for the Deaf and Hard of Hearing population has been extremely rewarding,” says MITRE’s Mike Woodman, who leads MITRE’s ACE Quill team.
Setting a New Standard for Effective Communication
This speech recognition work opened a new avenue of research. “It led us to a deeper dive on what constitutes ‘effective communication,’” Singh says.
The result was MITRE’s Effective Communications Evaluation Guide, which has become a resource for the FCC, Gallaudet University (GU), Rochester Institute of Technology National Technical Institute for the Deaf (RIT-NTID), and additional stakeholders. It outlines best practices, methodologies, and measures that contribute to achieving effective and equitable communication.
Our collaboration with the FCC in other areas continued, including development and enhancements to the open source, omni-channel platform called ACE Direct™ that debuted in 2016. We built on the capabilities provided by the ASR engine prototype ACE Quill to introduce captioning to ACE Direct, which was originally designed to give deaf users the ability to communicate one-on-one in American Sign Language. The enhancements expanded ACE Direct’s audience to include the Hard of Hearing population.
“We’ve continued to add user-centric features and new capabilities. An upcoming release includes language translation (11 languages) and transcription capabilities. Accurate translation and transcription capabilities are a game changer that will address an even wider audience of communications challenges faced by the general public,” says Dwight Handon, MITRE deputy project lead.
ACE Direct is a true “omni-channel” solution that can be used in any number of ways. For example, it’s an effective telemedicine platform, allowing for safe, timely access to healthcare services for expanded patient populations. The system can also improve conversation privacy, system security, and effectiveness of communications for services such as government agencies, banking, emergency 911, social services, language translation, customer service, and more.
Robust Partnerships Integral to Success
None of this would have been possible without the ongoing partnerships among the FCC, MITRE, GU, RIT-NTID, Hearing Loss Association of America, and the DHOH community.
Furthermore, MITRE has developed deep relationships with GU and RIT-NTID, performing VRS interoperability testing and exploring innovative solutions to complex telecommunication problems that affect TRS users.
Through our comprehensive, independent, and unbiased interoperability testing, we’ve significantly reduced, although not completely eliminated, the number-one video relay service consumer complaint, shifting the focus from “I am having problems completing a call” to “What other features can we add to the service to improve my outcomes?”
MITRE has also benefitted from the support of over a dozen interns from GU and RIT-NTID working with the FCC National Test Lab. The interns’ contributions helped us fine-tune several of the technologies we developed. Plus, they gave us a firsthand look at some of the ongoing technical challenges users who are deaf face every day.
“Since the beginning, we’ve worked closely with the FCC to apply our talent and technological resources to this problem in a holistic manner with a focus on the agency’s mission,” Singh says.
“It aligns with MITRE’s long-standing commitment to solving tough problems—in this case, one that existed within an underserved community.”
—by Kay M. Upham
Explore More at MITRE Focal Point: Promoting Equity