Distance-based time series classification approach for task recognition with application in surgical robot autonomy.

Area of Science:

• Robotics
• Surgical Technology
• Artificial Intelligence

Background:

• Robotic-assisted surgery offers enhanced precision but requires manual surgeon control.
• Current systems lack automatic task recognition, limiting adaptive adjustments like camera positioning.
• Automating task identification is crucial for developing more responsive robotic surgical systems.

Purpose of the Study:

• To develop an automated system for recognizing surgical tasks and gestures in robotic-assisted surgery.
• To enhance the responsiveness and adaptiveness of robotic surgical systems.

Main Methods:

• A novel distance-based time series classification framework was developed.
• Dynamic time warping distance was used to analyze temporal trajectory data of robot arms.
• A k-nearest neighbor algorithm was employed for surgical task and gesture classification.

Main Results:

• The proposed framework demonstrated superior performance compared to existing methods.
• Outperformed state-of-the-art methods by up to 9% in classifying surgical tasks.
• Achieved an 8% improvement in classifying surgical gestures.

Conclusions:

• The developed framework is robust and accurate for identifying surgeon actions.
• Enables the creation of adaptive control systems for robotic surgery.
• Facilitates more intuitive and responsive surgeon-robot interaction by predicting surgeon movements.