Development of a Vision-Guided Shared-Control System for Assistive Robotic Manipulators

Area of Science:

• Robotics and Human-Computer Interaction
• Assistive Technology
• Rehabilitation Engineering

Background:

• Upper-limb impairments significantly reduce independence in daily tasks.
• Controlling multi-degree-of-freedom assistive robotic manipulators (ARMs) presents a significant challenge.
• Existing control methods may not adequately address the complexities of ARM operation for users with disabilities.

Purpose of the Study:

• To develop and evaluate a novel vision-guided shared-control (VGS) system for ARMs.
• To assess the effectiveness of VGS in improving task performance for individuals with manipulation deficits.
• To compare VGS control against traditional teleoperation for a multi-step drinking task.

Main Methods:

• Development of a VGS system integrating user teleoperation for gross motion and autonomous control for fine manipulation.
• Bench-top testing of autonomous subtask success rates, achieving 80%–100%.
• Evaluation with three test pilots and one participant with spinal cord injury, measuring task success, completion time, and workload.

Main Results:

• VGS significantly improved overall task performance compared to teleoperation, including higher success rates and reduced task completion times.
• The VGS system demonstrated high success rates (80-100%) in autonomous subtasks.
• The spinal cord injury participant showed improved performance and reduced perceived workload with VGS, though usability ratings were similar to teleoperation.

Conclusions:

• The vision-guided shared-control (VGS) system offers a viable approach to enhance ARM control for individuals with upper-limb impairments.
• VGS effectively combines user intent with autonomous capabilities, leading to better task outcomes.
• Further research and testing are necessary to optimize VGS for broader application in assistive robotics for people with disabilities.