Lessons Learned from Designing and Evaluating a Robot-assisted Feeding System for Out-of-lab Use

Area of Science:

• Robotics
• Human-Computer Interaction
• Assistive Technology

Background:

• Millions struggle with independent eating due to disabilities, leading to social and emotional challenges during caregiver-assisted meals.
• Existing robot-assisted feeding research often focuses on isolated components in lab settings, neglecting real-world application.
• There is a critical need for end-to-end robot feeding systems evaluated in diverse, out-of-lab environments.

Purpose of the Study:

• To develop and evaluate an end-to-end robot-assisted feeding system for individuals with motor impairments.
• To address the challenges of out-of-lab deployment, focusing on user customizability and control for off-nominal situations.
• To empower individuals with motor impairments to feed themselves independently across various environments.

Main Methods:

• Collaborative development of a robot feeding system with community researchers (CRs) with motor impairments.
• Implementation of a user interface with substantial customizability and control.
• User-in-the-loop bite selection for food generalization and portable hardware for diverse environments.
• Two studies involving users with motor impairments and CRs feeding themselves in real-world settings (cafeteria, office, home).

Main Results:

• Study 1: Five users with motor impairments and one CR successfully used the system in various out-of-lab settings.
• Study 2: One CR used the system for five days in his home, feeding himself 10 meals across diverse contexts.
• Key lessons learned: spatial context necessitates customizability, variable autonomy aids in overcoming off-nominal events, and assistive robot benefits are context-dependent.

Conclusions:

• Customizability is crucial for users to adapt robot feeding systems to diverse spatial contexts.
• Variable autonomy empowers users to manage inevitable off-nominal events during robot-assisted feeding.
• The effectiveness and benefits of assistive robots are intrinsically linked to the specific context of use.