Robot-Assisted Group Tours for Blind People

TL;DR

This work investigates how a mobile robot can enable blind users to participate in mixed-visual guided group tours in public spaces. Through a three-phase study—interviews with blind participants and science communicators, system design and implementation, and field evaluation in a science museum—the authors identify key needs in communication, environmental awareness, and independent navigation, then implement a robot with environment description, guide/nearby-people reporting, and group-following capabilities. Field results show high usability and positive experiences, with the robot enhancing safety and engagement while also introducing potential distractions and self-consciousness, underscoring the need for personalization and shared autonomy. The study contributes design guidelines, a working robotic prototype, and empirical evidence that robot-assisted group tours can broaden accessibility and social participation for blind visitors in mixed-visual group contexts.

Abstract

Group interactions are essential to social functioning, yet effective engagement relies on the ability to recognize and interpret visual cues, making such engagement a significant challenge for blind people. In this paper, we investigate how a mobile robot can support group interactions for blind people. We used the scenario of a guided tour with mixed-visual groups involving blind and sighted visitors. Based on insights from an interview study with blind people (n=5) and museum experts (n=5), we designed and prototyped a robotic system that supported blind visitors to join group tours. We conducted a field study in a science museum where each blind participant (n=8) joined a group tour with one guide and two sighted participants (n=8). Findings indicated users' sense of safety from the robot's navigational support, concerns in the group participation, and preferences for obtaining environmental information. We present design implications for future robotic systems to support blind people's mixed-visual group participation.

Figures (4)

• Figure 1: Study process overview. Left: We first conducted interviews with both blind participants and museum experts, i.e., science communicators (SCs), who have experience with accessibility projects. Based on the reported challenges and opportunities, we developed a robotic system to aid group tour participation. Right: We evaluated the tour experience with the robot by recruiting blind participants, sighted tour visitors, and a tour guide. For each blind participant, a training session was held with the guide before the tour.
• Figure 2: System Overview. The blind user presses the buttons on the robot handle and hears responses via a phone. The robot has five main features to support the blind visitor's group tours: 1) The robot uses its cameras to send images and its current position to the LLM server, and receives a surrounding description; 2) The robot's phone receives UWB broadcasting signals from the phone carried by the guide, and the robot determines the location of the guide; 3) The robot uses its cameras and a local detection model mmdetection to detect nearby people; 4) The robot triggers a notification on the phone carried by the guide; and 5) The robot navigates via tele-operation while using its LiDAR for collision avoidance.
• Figure 3: Tour timelines with 1) tour stages, 2) robot navigational behaviors, and 3) features triggered by users. On the tour, the group visited a fixed set of exhibits (solid line) guided by the tour guide and had touch experiences (dotted line). The space in between the lines represents the transition from one exhibit to the next. The user used the robot to follow the robot (gray block) and can get stuck sometimes due to crowds or obstacles (dark gray block). The timeline also illustrated the four features triggered by the user before, during, and after the tour.
• Figure 4: Left: How many times each feature was triggered during the tours experienced by participants? The mean and SD of each feature usage are: Surrounding Description (Mean 4.13, SD 2.10), Guide Location (Mean 4.88, SD 2.70), Nearby People Locations (Mean 2.75, SD 2.19), Call Guide (Mean 0.75, SD 1.16). Additionally, the stop feature, which mutes the robot, was triggered with a mean of 7.62 and a SD of 6.20. As reported by the participants, because there is a slight delay between triggering this feature and muting the robot, they would often trigger it repeatedly, resulting in a high count. Right: Percentage of tour stages when participants used the robot's features.