Bio-behavioral Team Dynamics Measurement System: Multimodal Sensing, Dynamical Systems Modeling, and Machine Learning...

Area of Science:

• Human-computer interaction
• Team dynamics
• Biometric analysis

Background:

• Military training requires objective team performance assessment.
• Current methods are often subjective and time-consuming.
• DARPA's OP TEMPO program aims to enhance warfighter readiness through automated assessment.

Purpose of the Study:

• To develop and validate the Bio-behavioral Team Dynamics Measurement System (BioTDMS).
• To create an objective, automatic system for assessing team performance in simulation training.
• To identify bio-behavioral signatures that predict team effectiveness.

Main Methods:

• Utilized a multimodal sensor suite to collect neural, cardio-respiratory, eye tracking, and verbal data.
• Employed a layered symbolic dynamics model and moving-window entropy/mutual information for real-time analysis.
• Integrated a multitask, multi-kernel learning engine for performance prediction and explainability.

Main Results:

• BioTDMS achieved 90% accuracy in predicting subjective team performance assessments.
• Demonstrated real-time metrics for team adaptability and influence distribution.
• Successfully implemented and field-tested during U.S. Marine Corps Fire Support Team (FiST) training.

Conclusions:

• BioTDMS shows significant potential as an operational tool for objective team assessments.
• The system can enhance training effectiveness by providing immediate, data-driven feedback.
• Future work will explore generalizability in air combat teams and human-autonomy teaming.