Toward Dynamically Adaptive Simulation: Multimodal Classification of User Expertise Using Wearable Devices.

Area of Science:

• Biomedical Engineering
• Educational Technology
• Human-Computer Interaction

Background:

• Simulation-based training is effective but learning suffers if simulation difficulty mismatches learner expertise.
• Adaptive learning systems require accurate assessment of a learner's skill level.
• Wearable sensors offer a non-intrusive method for monitoring physiological responses.

Purpose of the Study:

• To propose a novel framework for adaptive simulation that dynamically adjusts complexity based on learner expertise.
• To investigate the use of biological signals for classifying learner expertise in trauma response simulations.
• To develop a system that matches simulation difficulty to individual learner capabilities for optimized learning.

Main Methods:

• Collected electrocardiogram (ECG) and galvanic skin response (GSR) data from novice and expert trauma responders during simulations.
• Applied feature extraction and selection techniques to the collected biological signals.
• Utilized various machine learning algorithms for multimodal expertise classification based on bio-signals.

Main Results:

• Demonstrated the feasibility of classifying learner expertise using a combination of ECG and GSR signals.
• Achieved successful multimodal classification of trauma responders' expertise levels.
• Validated the potential of bio-signal analysis for adaptive simulation applications.

Conclusions:

• Biological signals, specifically ECG and GSR, can be effectively used for classifying learner expertise.
• This multimodal bio-signal classification approach supports the development of adaptive simulation frameworks.
• Optimizing simulation difficulty through adaptive systems enhances learning outcomes in critical training scenarios.