Designing Accessible Robot Communication for Blind People

Mina Huh, Huihan Liu, Albert Yu, Roberto Martín-Martín, Yuke Zhu, Maya Cakmak, Amy Pavel

Keywords: Human-Robot Interaction, Accessibility, Transparency

TL;DR: Through two user studies (in-person and online controlled study, N=50), we characterize the strategies and challenges of monitoring household robots for blind people and derive design guidelines for accessible robot communication.

Abstract:

Robots are moving into homes with the promise to reduce barriers to housework for people with disabilities and decrease effort for
everyone. As robots perform tasks autonomously, people need to monitor their progress to verify task execution and intervene when
necessary. While sighted users can visually observe the robot, blind users lack timely access to the robot’s actions and task outcomes.
Our work investigates strategies and challenges that blind people encounter when monitoring robots to inform how to make robot
task communication accessible. To understand how blind people monitor robots, we conducted an in-person observational study with
10 blind participants using two robot platforms. Participants primarily used non-visual cues — listening for actions during execution
and using touch to inspect the workspace afterward. However, these cues were often ambiguous, leading to missed robot errors and
uncertainty about task outcomes. Some participants also turned to visual interpretation tools (e.g., BeMyEyes, Meta Glasses), but
these tools produced generic scene descriptions rather than descriptions of the robot’s task progress. Participants therefore requested
proactive, task-relevant robot narration, and the ability to ask questions. To study these needs at scale, we developed an interactive,
AI-powered communication prototype that supports voice-based robot narration and question answering. Our study with 20 blind and
20 sighted participants reveals disparities in question-asking strategies and preferences across groups. We distill design guidelines for
accessible robot communication that improves transparency for blind people.

Submission Number: 6