Milo, a Fully Autonomous Indoor/Outdoor Robotic Guide Dog

2026-07-21

AI Summary

This work addresses the high cost, long waiting periods, and limited lifespan of traditional guide dogs by proposing Milo—a low-cost (approximately $2,000), fully autonomous robotic guide dog. Built upon the Unitree Go2 platform, Milo operates without prior environmental knowledge and performs all computation onboard. It integrates a voxel-based mapping system with dedicated perception modules for terrain, obstacles, and pedestrians, and employs an obstacle-avoidance policy trained in bird’s-eye-view simulation to enable seamless indoor–outdoor navigation. As the first open-source, general-purpose robotic guide dog system, Milo demonstrates superior navigation smoothness and a lower user collision rate compared to conventional costmap-based baselines in real-world, complex environments.

Abstract

Many Blind and Low-Vision (BLV) people rely on guide dogs for moment-to-moment navigation, such as staying on path and avoiding obstacles and pedestrians. However, guide dogs are expensive to acquire and maintain (approximately \$50k USD plus ongoing costs), often involve long waiting lists, and have relatively short life expectancies. While robot guide dogs offer a promising alternative, existing approaches exploring this idea suffer from several drawbacks: They often lack the autonomy required for real-world deployment, relying on prior 3D scans of the environment, external computation, or limited awareness of the handler. In this work, we present Milo, the first open-source, low-cost (approximately \$2k USD) robotic guide dog platform capable of fulfilling the basic collaborative navigation role expected of a guide dog. Milo is fully autonomous, requiring no a priori knowledge of the environment, completely self-contained with all computation performed onboard, and suitable for both indoor and outdoor navigation while avoiding obstacles and pedestrians. Our system consists of a modified Unitree Go2 robot (equipped with onboard compute, sensors, and a handle), a perception stack combining voxel mapping with floor, obstacle, and pedestrian detection, and a navigation stack based on an obstacle-avoidance policy trained in a custom bird's-eye-view simulator. We evaluate Milo in real indoor and outdoor obstacle courses and compare it against a costmap-based baseline, demonstrating smoother navigation and fewer handler collisions. To maximize accessibility for BLV users, we release both the robot hardware instructions and the complete software stack as open source.

Problem

• robotic guide dog
• autonomous navigation
• blind and low-vision assistance
• obstacle avoidance
• indoor/outdoor mobility

Innovation

• autonomous robotic guide dog
• onboard perception and navigation
• voxel-based mapping
• simulation-trained obstacle avoidance
• open-source assistive robotics