A video processing and machine learning based method for evaluating safety-critical operator engagement in a motorway...

Area of Science:

• Human-Computer Interaction
• Artificial Intelligence
• Computer Vision

Background:

• Operator engagement is crucial for safety in critical systems.
• Existing engagement measurement methods have limitations in real-world settings.
• Automated, objective engagement detection is needed.

Purpose of the Study:

• To develop and validate a novel AI-driven methodology for evaluating operator engagement.
• To address limitations of current engagement measurement techniques.
• To enhance safety in automatic systems through better engagement monitoring.

Main Methods:

• Utilized Artificial Intelligence (AI) with computer vision (OpenCV, OpenPose) to analyze operator body posture.
• Developed a Support Vector Machine (SVM) model for engagement state classification.
• Collected data from motorway control room operators.

Main Results:

• Achieved an average accuracy of 0.89 in engagement evaluation.
• Reported weighted average precision, recall, and F1-scores above 0.84.
• Demonstrated the effectiveness of the AI-based framework.

Conclusions:

• The proposed AI methodology provides an effective, real-time, and objective measure of operator engagement.
• Accurate engagement measurement can inform interventions to improve operator performance and system safety.
• Highlights the importance of data labeling for engagement state measurement.