Carnegie Mellon Team Develops Disability-Friendly Devices

 Initiative to Develop Cost-Effective Assistive Technologies for the Blind Using Current Technology

May 8, 2006

Sometimes tech projects that make the largest impact don't revolve around high-end, new technology. The key is the approach taken to finding a solution for the problem at hand, and being able to combine current technology with cutting-edge tech. Take Trinetra, for instance, a project that aims to develop cost-effective assistive technologies to provide blind people with a greater degree of independence in their daily activities. The Sanskrit word Trinetra refers to the powerful third eye of the Hindu god Shiva.

Trinetra was started by Priya Narasimhan, an assistant professor of electrical and computer engineering and computer science, who struck upon the idea for the project one day in December 2004 when she noticed how difficult it was for blind people to catch a bus on a snowy Pittsburgh evening. Patrick Lanigan, a graduate student in the Information Networking Institute (INI), Aaron Paulos, a Research Programmer in ECE, and Andrew Williams, an ECE graduate student, all opted to work under Narasimhan for Trinetra because of the passion they shared for working on a project that had the potential to help the significant blind population of Pittsburgh.

Trinetra technology has just been installed in Carnegie Mellon's campus store, Entropy, to make it easier for blind people to go grocery shopping. This assistive technology works as follows: imagine a blind man walks into Entropy and wants to pick up a bottle of mustard. As he walks past one aisle to another, he uses a UPC-reading Baracoda pencil integrated with his phone to scan the barcodes on the shelves under the products he browses. The Baracoda pencil contacts the UPC database through the Internet-enabled phone and identifies the product. The information returned to the phone is then read out by TALKS, software developed by Nuance that was installed in the phone.

Dan Rossi, mentor to the Trinetra team, has been blind since he was seven; he is also a database system administrator. A highly accomplished person in the area of technology, he was the guiding force for Trinetra, with researchers frequently waiting for his opinion on certain aspects of the project before moving ahead.

"We have involved a blind person from day one in the design of the project so that blind people's needs and inputs have been factored into our design in an intrinsic and fundamental way," said Narasimhan.

"You cannot imagine the real impact of this," Rossi said. "When shopping with a store assistant, it is nearly impossible to browse products. I ask for what I need, and they take me to that product, and that's it. A device that can tell me what just about anything in the store actually is is incredible."

There were many challenges faced by the Trinetra project. "The major challenge for us is to keep the project's outcomes cost-effective. If we develop technologies that end up increasing a blind person's cost of living significantly, we have really not lived up to our mission," Narasimhan said. "The other challenge is to resist the temptation to add in bells and whistles which, while they might be attractive to look at, have limited practical use to blind people." Since devices for the visually disabled are a niche market, they translate into high costs for the end user. Trinetra's bag of gadgets includes mostly off-the-shelf items that people without disabilities use every day.

One of the unique features of Trinetra that sets it apart from other ongoing projects is that it leverages available infrastructure. The developers don't have to pay for the barcode (UPC) database; it is already available on the Internet. These design decisions help keep costs down for the project, so insufficient funds have not been a cause of worry so far for the Trinetra team. When the project began, Narasimhan took money off her payroll to keep the project going. Since then, the project has survived comfortably on the $7500 of funding it received from the Pennsylvania Cyber Security Commercialization Initiative (PACSCI).

But there's more to worry about than just keeping costs low in a project like this. Narasimhan pointed out the privacy and safety issues that might arise for people using these services. Such issues and other more intricate ones have to be understood and analyzed simultaneously with the development of assistive technologies.

Trinetra's next mission is to make Carnegie Mellon's shuttle service friendlier to the visually impaired by keeping them informed of arrival times. The phone is again the preferred medium of communication, this time using text messages. To start things off, Narasimhan had her research students go around and try to get a feel of the shuttle service. The project might also end up being useful for students using the service late at night.

So what keeps the Trinetra team going? "The team designed, developed, and tested a concept that not only works, but is completely useful," said Rossi. "This is not a sensor net vest of pager motors and scanning hats that are a neat idea, but not practical in the least. The Trinetra system is practical and useful and I commend the team on their work."

The future looks bright for the Trinetra team. Narasimhan said, "My aim is for Trinetra to become one of the research thrusts of my research group, so that we can explore the application of embedded, distributed technologies to improve the quality of life for disabled people."

Reprinted with permission from The Tartan Online, April 10, 2006. Text by Saravana Sivasankaran.

Left to right: Andrew Williams, Aaron Paulos, Patrick Lanigan, and the project's leader, Priya Narasimhan, show their research.

Dan Rossi, the project's mentor and a user of Trinetra, scans a product.

Related People:

Priya Narasimhan