Engineers create a robotic eye

Cost factor

The limited use of real dogs for guidance is primarily down to its cost of around $50,000, which requires two to three years to train. Furthermore, only around half of the canines that complete their training go on to service visually impaired persons. Seeing-eye robot dogs have the potential to significantly reduce costs, enhance efficiency, and increase accessibility, according to the team.

The effort by the team at Binghamton University is one of the first attempts to create a seeing-eye robot after the discovery and cost reduction of quadruped technology. After nearly a year of development, the team developed a novel leash-tugging interface for use with reinforcement learning.

“In about 10 hours of training, these robots are able to move around, navigating the indoor environment, guiding people, avoiding obstacles, and at the same time, being able to detect the tugs,” said Zhang.

The pulling interface allows the user to pull the robot in a certain direction at a hallway intersection, causing the robot to turn in response. While the robot shows potential, engineers believe that further research and development is required before the technology is suitable for use in specific contexts.

Language interface

Next on the agenda is the addition of a natural language interface. The ability could help a visually impaired person chat with the robot to obtain some assistance, which furthers the accessibility factor. Engineers say that intelligent disobedience is also a helpful skill. "For example, if I’m visually impaired and I tell the robot dog to walk into traffic, we would want the robot to understand that. We should disregard what the human wants in that situation. Those are some future directions we’re looking into," said the team.

Researchers are also in touch with organizations that help them acquire direct and valuable feedback from the visually impaired community. Suggestion regarding the danger of sudden drop-offs in pathways

While the team is not restricting the technology's capabilities, their feedback and intuition lead them to believe the robots may be more beneficial in some circumstances. Because the robots can carry maps of particularly difficult-to-navigate areas, they may be more effective than genuine seeing-eye dogs at guiding visually impaired persons to their preferred destinations.

While the research is still in its early phases, the team feels it is a promising step towards expanding public area accessibility for the visually impaired community. "If this is going well, then potentially in a few years we can set up this seeing-eye robot dog at shopping malls and airports. It’s pretty much like how people use shared bicycles on campus,” said Zhang.