What features are essential for a robotic guide dog? According to researchers, the best way to find out is to ask the blind. A recent study, led by the University of Massachusetts Amherst, has identified the key features needed to develop robot guide dogs based on insights from guide dog users and trainers. This study was honored with a Best Paper Award at CHI 2024: Conference on Human Factors in Computing Systems (CHI). Guide dogs play a crucial role in providing independence and mobility for the visually impaired.Guidedogs are a valuable resource for individuals with visual impairments, but only a small percentage of those individuals have access to them due to various barriers such as the limited availability of trained dogs, high costs, allergies, and other physical limitations. Robots have the potential to fill this gap and meet the needs of those who can’t have a guide dog if designers can develop the right features. Donghyun Kim, assistant professor at UMass Amherst Manning College of Information and Computer Science, highlights the potential of guide-dog robots in addressing this pressing need.The paper, which has won awards, emphasizes the importance of understanding the needs of visually impaired individuals who rely on guide dogs before developing robotic technology to assist them. The research team conducted interviews and observation sessions with 23 visually impaired dog-guide handlers and five trainers to identify the current limitations of guide dogs and the qualities that handlers are seeking in an effective guide. This information will be crucial in the development of future robotic guide dogs.One of the more complex themes that emerged from the interviews was the delicate balance between robot autonomy and human control. According to Kim, “Originally, we thought that we were developing an autonomous driving car.” The vision was that the user would instruct the robot where to go, and the robot would then navigate independently to that destination with the user in tow.
However, this turned out not to be the case.
The interviews uncovered that handlers do not use their dog as a global navigation system. Instead, the handler controls the overall route while the dog is responsible for local obstacle avoidance. However, even though theIt’s not a strict rule. Dogs can also learn familiar paths and may eventually lead a person to regular destinations without needing directions from the handler.
“When the handler trusts the dog and gives the dog more freedom, it’s a bit delicate,” says Kim. “We can’t just create a robot that is completely passive, only following the handler, or completely independent, because then the handler feels unsafe.”
The researchers hope this article will be a helpful guide, not only in Kim’s lab, but for other robot developers as well. “In this article, we also provide guidance on how we should develop these robots to make them behave autonomously while maintaining safety for the handler.”Hochul Hwang, a doctoral candidate in Kim’s robotics lab and the first author of the paper, emphasizes the importance of making the robots deployable in real-world scenarios. He notes that a two-hour battery life is crucial for commuting, as it covers a significant portion of the average commute time. In fact, about 90% of the people surveyed mentioned the battery life as a key factor. Hwang explains that designing hardware with a longer battery life is critical, as current quadruped robots do not last for two hours.
The paper also highlights other important findings, such as the need to add more camera orientations to help the robots navigate overhead obstacles and the addition of audio sensors to detect hazards approaching from obscured areas.The most exciting aspect of winning this award is that the research community recognizes and understands ‘sidewalk’ to convey the cue, “go straight,” which means follow the street (not travel in a perfectly straight line); and helping users get on the right bus (and then find a seat as well). This paper is a great starting point, adding that there is even more information to unpack from their 2,000 minutes of audio and 240 minutes of video data. Winning the Best Paper Award was a distinction that put the work in the top 1% of all papers submitted to the conference.”Kim emphasizes the importance of our mission,” states Kim. “Given that we do not anticipate the availability of guide dog robots for visually impaired individuals within the next year, and acknowledging that we may not solve every problem, we aim to inspire a wide range of robotics and human-robot interaction researchers with this paper, in the hopes of realizing our vision sooner.”
The paper also includes contributions from other researchers, including:
Ivan Lee, associate professor at CICS and co-corresponding author of the article alongside Donghyun, an expert in adaptive technologies and human-centered design; Joydeep Biswas, associate professor at the University of Texas Austin, who contributed his expertise in the field.experience in developing artificial intelligence (AI) algorithms for enabling robots to maneuver through unorganized surroundings; Hee Tae Jung, assistant professor at Indiana University, who contributed his knowledge in human factors and qualitative research to an inclusive study with individuals with chronic illnesses; and Nicholas Giudice, a professor at the University of Maine who is visually impaired and offered valuable perspectives and analysis of the interviews.
In the end, Kim recognizes that robotics can have the greatest impact when scientists consider the human aspect. “My Ph.D. and postdoctoral research is focused on finding ways to optimize the functionality of these robots while prioritizing the human experience.”Reaching the C-H-A-I-R_L-A-B team’s goal of developing a guide dog robot that can support visually impaired people involves understanding the interactions between guide dogs and their handlers. To achieve this, researchers have studied and observed the dynamics between handler-guide dog duos, and will apply these insights to the design of the guide dog robot. This project aims to create a practical and meaningful application that will benefit humanity. The findings of this research have been published in the CHI 2024: Conference on Human Factors in Computing Systems.81
