Giudice part of team behind award-winning paper on robotic guide dogs

University of Maine professor of spatial computing Nick Giudice played a key role as a contributor to an award-winning paper about developing robot guide dogs to assist the visually impaired.

The paper – which is titled “Toward Robotic Companions: Understanding Handler-Guide Dog Interactions for Informed Guide Dog Robot Design” and can be found here – came from a study led by researchers at the University of Massachusetts Amherst. It won a Best Paper Award at CHI 2024: Conference on Human Factors in Computing Systems, the leading venue for human-computer interaction research. You can find the full UMass Amherst story here.

While Giudice, who is also the founder of and chief research scientist at UMaine’s VEMI Lab, was not a primary investigator on this project, his combination of expertise and experience make him an ideal fit for this project.

“I have spent a lot of my career working on new tech for improving spatial awareness, navigation, and information access for blind and low vision people. I have also been a guide dog user for over 30 years,” said Giudice. “The combination of these personal and professional interests gives me a perspective on developing smart guiding technology that most people simply do not have.

“This lived experience provides me a lot of insight of what tech works and what doesn’t and what problems truly exist when navigating for blind folks – insight that is critical for designing solutions that avoid what I call the ‘engineering trap’ and creating tools that actually make a difference.”

When the PI of the team, assistant professor in UMass Amherst’s College of Information and Computer Science Donghyun Kim, reached out for input on this paper, Giudice embraced the opportunity, though he’s historically been somewhat skeptical in this particular area.

“[Kim] knew of my work in the area of assistive technology supporting navigation for blind individuals and wanted my input on their project,” he said. “I have been aware of robotic guide dogs for years but have always had a healthy skepticism, as the work has traditionally been driven by engineering factors rather than an understanding of the human-dog relationship.

“However, this group has a very different approach, starting their development by trying to understand the handler- animal dog relationship and how this could be accurately modeled in the human-robotic dog team,” he continued. “Their approach of interviewing blind guide dog users and guide dog trainers, as well as observing working teams and even trying working under blindfold was impressive to me. Their approach convinced me to get involved in the project and it ultimately led to a very interesting and I hope influential paper in the field for other developers and designers.”

According to Giudice, this paper provides the first user-driven guidelines on the needs and design specifications for robot guide dogs. He believes there’s tremendous potential value to the field and to creating new mobility aids that can complement (but not replace) current biological dogs.

“In my head, the robotic guide dog is not a replacement to an animal guide, it is a complement to be used in situations where I don’t want to take my dog but still want similar guidance,” he said. “For instance, in situations where it is not safe, such as if there is glass on the ground or if it is too hot or cold outside,  if it is too loud (like at fireworks displays or concerts) or too busy (such as at a crowded bar or sporting event). These are places where I wouldn’t want to take my ‘real’ guide dog but where a robotic guide dog would be fine.

“For all of these reasons, I am now completely on board with robotic guides as a potential solution, although it’s worth noting that a LOT of work is still needed for them to be truly effective. This work starts with talking to blind users and this project starts that process.”

Giudice stresses that this paper is just the first step toward true understanding of user needs and specifications – a starting point for the next era of research in robotic guide dog technology.

“I believe that given the fast-paced evolution in computational power, sensor technology, and embedded AI-based processing of new legged robots, what is difficult now will increasingly be possible in the next few years,” he said “This exponential tech development, coupled with more robust understanding of user needs and the handler-dog team, will allow us to design future guide robots that provide a viable alternative to complement existing animal dogs.”

Giudice went on to add that he believes that the eventual acceptance and adoption of this technology will hinge on how well projects such as this one try to understand current guide teams and create solutions that add new value and information access.