GuideNav wins a Systems Paper Honorable Mention at HRI 2026

Our paper on GuideNav received a Systems Paper Honorable Mention at HRI 2026.

GuideNav is a vision-only, teach-and-repeat navigation system inspired by how guide dogs are trained: a sighted person walks a route once, and the robot repeats it autonomously. In experiments it followed kilometer-scale routes across five outdoor environments.

The design was grounded in extensive formative research — 40 interviews (26 guide dog handlers, 4 white cane users, 9 guide dog trainers, 1 O&M specialist) and 15+ hours of real-world observation. The resulting Handler–Guide-dog Interaction (HGI) dataset is publicly available.

Congratulations to Hochul Hwang for leading this effort, and to co-authors Soowan Yang, Jahir Sadik Monon, Nicholas Giudice, Sunghoon Ivan Lee, and Joydeep Biswas.

• Project website: https://guidedogrobot-navigation.github.io/
• HGI dataset: https://guidedogrobot-hgidataset.github.io/