Beyond Omakase: Designing Shared Control for Navigation Robots with Blind People

Research Background and Problem

• Problem or Challenge: The authors point out that while autonomous navigation robots can enhance the independence of blind individuals, existing designs often adopt an "omakase" model—leaving all decision-making to the robot. This approach may undermine user control, foster dependency, and limit effective interaction in socially complex environments.
• Significance: Blind individuals often face complex social environments during navigation, including moving crowds, queues, and crowded spaces. Current navigation assistive technologies fail to adequately address these challenges and do not provide sufficient user control, negatively impacting their independence, comfort, and social acceptance.
• Research Motivation and Related Work: Literature indicates that enhancing user control over navigation systems can increase trust and comfort while addressing the limited adaptability of navigation technologies to social norms. This study aims to explore interaction designs for robots suitable for blind users in complex social environments.

Solution

• Proposed Approach: This study designs a shared control navigation method, incorporating three user control modes: the Omakase mode (robot takes full control), the Monitor mode (user monitors environmental information), and the Boss mode (user actively controls the robot).
• Innovation: The primary innovation lies in the design of a multi-mode shared control framework, allowing blind users to flexibly switch modes based on the environment. This design considers users' perception, decision-making abilities, and social navigation needs, balancing robot support and user autonomy. Additionally, the study integrates blind users' social navigation experiences to refine the design of these three modes.
• Implementation Steps and Key Techniques:
  1. Preliminary Research: Conduct structured interviews with 14 blind individuals to gain in-depth insights into challenges in social navigation (e.g., moving crowds, queuing issues) and coping strategies.
  2. User Experimentation: Engage 13 participants in real-world environments to experience navigation tasks with the robot, testing the effectiveness of the three control modes.
  3. Follow-Up Interviews and Data Analysis: Collect participant feedback on the modes and analyze how they chose and switched between modes in different environments.

Research Findings

• Specific Findings:
  • Participants preferred the Monitor mode for environmental monitoring and the Boss mode for actively managing interactions between the robot and people.
  • In complex social environments, a collaborative dynamic between the user and the robot was achieved: the robot provided informational assistance, while the user made key decisions.
  • Detailed feedback on functional requirements included users' desire for the robot to provide information about the environment, nearby people, and navigation goals.
• Advantages Over Existing Solutions:
  • Enhanced user autonomy and environmental awareness, addressing the issue of complete reliance on machines in the traditional "omakase" model.
  • Offered highly flexible mode-switching for social navigation, aiding adaptation to various environmental conditions.
• Experimental or Evaluation Results:
  • The Monitor mode was frequently used for environmental assessment, indicating its importance in providing critical informational support.
  • The Boss mode played a central role in user-controlled navigation decisions and effectively facilitated interaction in social scenarios.
  • Task performance in dynamic crowd and queuing scenarios demonstrated that the shared control approach helps users better engage with and resolve navigation challenges.
• Limitations and Future Directions:
  • Current experimental scenarios were limited to a museum environment; further research should expand to more complex settings such as train stations or shopping areas.
  • The Monitor mode employed a Wizard-of-Oz method (experimenter simulating robot responses); future work should develop reliable machine responses to enhance user trust.
  • Long-term studies could explore the evolution of interactions between blind users and robots, such as the need for users to develop personalized strategies for optimizing robot usage.

Conclusion

This study proposes a shared control framework to provide new pathways for enhancing the autonomy of blind individuals in complex navigation scenarios. Future research directions include technical implementation, expansion to more complex environments, and long-term behavioral studies with diverse user groups. The findings of this study contribute to guiding the design of future social navigation robots, fostering collaboration and trust between users and automated systems.