Operator functional state classification using least-square support vector machine based recursive feature...

Area of Science:

• Human-Computer Interaction
• Cognitive Engineering
• Biomedical Engineering

Background:

• Operator functional states (OFS) are crucial for safety in human-machine systems.
• Assessing OFS accurately is challenging due to complex physiological responses.
• Existing methods may lack generalization ability in dynamic environments.

Purpose of the Study:

• To propose novel psychophysiological-data-driven classification frameworks for OFS assessment.
• To develop robust methods for feature selection and classification of operator states.
• To ensure stable generalization ability in safety-critical applications.

Main Methods:

• Combined Recursive Feature Elimination (RFE) and Least Square Support Vector Machine (LSSVM) for feature selection.
• Developed binary and multiclass LSSVM classifiers for OFS assessment.
• Implemented multiclass LSSVM-RFE and Decision Directed Acyclic Graph (DDAG) for specific state recognition.

Main Results:

• Identified specific psychophysiological features characterizing different dimensions of OFS.
• Achieved high and stable classification accuracy using the proposed frameworks.
• Demonstrated the effectiveness of RFE in optimizing feature selection for OFS.

Conclusions:

• The proposed psychophysiological-data-driven frameworks offer stable OFS assessment capabilities.
• Proper implementation of RFE is key to achieving high classification accuracy.
• These methods enhance the reliability of human-machine systems in safety-critical scenarios.