Towards free 3D end-point control for robotic-assisted human reaching using binocular eye tracking

Area of Science:

• Biomedical Engineering
• Neuroscience
• Human-Robot Interaction

Background:

• Eye movements are direct behavioral indicators of action intentions and are preserved in paralysis.
• Existing eye-tracking interfaces for robotic control are limited by poor human-robot interaction.
• Movement disorders like stroke and Parkinson's disease often impair motor function but not eye control.

Purpose of the Study:

• To develop and evaluate a novel 3D gaze-tracking system for precise robotic arm control.
• To enable intuitive, eye-controlled interaction for patients with severe motor impairments.
• To overcome limitations of current human-robot interfaces in assistive technology.

Main Methods:

• Combined a GT3D binocular eye tracker with a custom 3D head tracking system.
• Developed a novel 3D calibration method using a space-filling Peano curve for automated calibration.
• User control involved looking at a target and winking, enabling end-point control of a robotic arm.

Main Results:

• Achieved continuous 3D end-point control of a robotic arm support system.
• Demonstrated high spatial accuracy with a Root Mean Square Error (RMSE) of 6 cm and a standard deviation of 4 cm.
• The automated calibration procedure yielded thousands of points, surpassing traditional methods.

Conclusions:

• The developed eye-tracking system offers a highly accurate and precise method for controlling robotic arms.
• This technology provides a viable assistive interface for individuals with movement disorders, retaining head mobility.
• The system represents a significant advancement in assistive robotics and human-computer interaction for impaired users.