Researchers at Binghamton University in New York have programmed a robotic seeing-eye dog to guide the visually impaired.

To develop the robotic seeing-eye dog, researchers created a leash tugging interface for training the robot.

Source: Binghamton University

Following 10 hours of training, the robotic seeing-eye dog was able to move around an indoor space, guide people, avoid obstacles and detect tugs.

Specifically, the tugging interface enables the user to turn the robot in a desired direction.

Going forward, the team expects to outfit the robot with a natural language interface that would allow for the user to have a conversation with the robot according to the situation. Further, the researchers expect to enable the robot to be intelligently disobedient in order to disregard dangerous instructions from the human such as walking into traffic.

The Binghamton team will present their research at the Conference on Robot Learning (CoRL), which is an annual international conference focusing on the intersection of robotics and machine learning. CoRL 2023 will be held in Atlanta, Georgia, from November 6 to 9, 2023.