Researchers unveil Milo, open-source robotic guide dog under $2,000

Researchers have developed Milo, described as the world's first fully autonomous robotic guide dog, built to serve as a lower-cost alternative to traditional guide dogs for blind and visually impaired people. Certified guide dogs in Canada can cost tens of thousands of dollars to train, and applicants often wait several years before receiving one; Milo could be manufactured for less than $2,000, significantly expanding access to robotic mobility assistance.

Unlike many autonomous robots that require pre-mapped environments, Milo is designed to navigate unfamiliar locations without a prior 3D scan. The robot uses onboard artificial intelligence to identify paths, avoid pedestrians and obstacles, and guide its user through indoor and outdoor spaces. It does not rely on an internet connection or cloud computing, instead performing all processing on board. The navigation system was trained using reinforcement learning inside a custom simulator that generated a wide variety of virtual environments and obstacles. Milo interprets its surroundings using a bird's-eye-view representation, enabling it to navigate consistently under different lighting and environmental conditions. The robot also builds and maintains a map of its surroundings as it travels, allowing it to remember obstacles that are no longer directly visible rather than reacting only to what its sensors currently detect.

To guide users safely, Milo tracks the handler's position using sensors integrated into a telescopic handle. A directional control pad allows users to adjust walking speed and indicate which direction they wish to travel at intersections. The developers evaluated the system on indoor and outdoor obstacle courses, measuring navigation accuracy, travel speed, and obstacle clearance. Alongside a research paper describing the project, the team has released Milo as an open-source platform, making its hardware designs, simulation software, and pre-trained AI models publicly available to encourage further research and development. The developers say the combination of lower cost, autonomous operation, and open-source availability could make robotic mobility assistance more accessible to people who are unable to obtain a traditional guide dog.