How Does Delegation in Social Interaction Evolve Over Time? Navigation with a Robot for Blind People

TL;DR

Blind users face navigation challenges that persist despite autonomous robotics. This work employs a three-week longitudinal study with a shared-control navigation robot, augmented by GPT-based environmental descriptions and obstacle explanations, to observe how delegation and collaboration evolve with repeated use. Key findings show that users progressively delegate social interactions, refine interpretation of robot behavior, and rely on contextual descriptions to calibrate actions, with design implications for adaptive, transparent, and multilingual interfaces. The study demonstrates that responsive, user-centered robots can balance autonomy with user agency in dynamic social environments, supporting long-term adoption in real-world settings.

Abstract

Autonomy and independent navigation are vital to daily life but remain challenging for individuals with blindness. Robotic systems can enhance mobility and confidence by providing intelligent navigation assistance. However, fully autonomous systems may reduce users' sense of control, even when they wish to remain actively involved. Although collaboration between user and robot has been recognized as important, little is known about how perceptions of this relationship change with repeated use. We present a repeated exposure study with six blind participants who interacted with a navigation-assistive robot in a real-world museum. Participants completed tasks such as navigating crowds, approaching lines, and encountering obstacles. Findings show that participants refined their strategies over time, developing clearer preferences about when to rely on the robot versus act independently. This work provides insights into how strategies and preferences evolve with repeated interaction and offers design implications for robots that adapt to user needs over time.

Figures (8)

• Figure 1: Guide Robot. The suitcase-shaped robot used in the study. It integrates a handle interface, smartphone, and external speaker for user interaction. It includes multimodal sensing through front and right RGBD cameras, a 360 LiDAR, a bottom LiDAR, and motorized wheels for navigation.
• Figure 2: Route Object Obstacles. Examples of object obstacles encountered during navigation with GPT-generated descriptions, as captured by the robot’s front RGBD camera. Week 1 included a green foam block from a children’s play exhibit, Week 2 featured two suitcases, and Week 3 presented a baby cart.
• Figure 3: Navigation Routes. Routes used in the study as mapped on museum floors. Week 1 and Week 3 were conducted on Floor A, while Week 2 was on Floor B. Across the routes, participants encountered a variety of obstacles, including crowds, lines, luggage, and baby carts.
• Figure 4: Ratings for Statement (1): Preference for the Robot to Ask for Help. Participant responses over three weeks showed a general upward trend, with P1, P4, and P5 exhibiting the strongest increases, indicating growing acceptance of the robot actively requesting assistance.
• Figure 5: Ratings for Statement (2): Preference for the Robot to Ask People to Move. Participant responses across three weeks showed steadily increasing agreement, with P1, P2, P4, and P5 exhibiting the largest gains, reflecting growing acceptance of robots directly requesting people to clear the way.