Adaptive and Embeddable Agents for Real-Time Cognitive Readiness and Performance

To meet the ever-increasing challenges imposed by next generation weapon systems and continued reduced manning efforts, Quantum Leap Innovations, Inc. will collaborate with the Institute for Simulation and Training (IST) at the University of Central Florida (UCF) to develop adaptive, embeddable, and configurable agents for real-time cognitive readiness and performance assessment. The objective of this SBIR is to develop and operationally test adaptive and embeddable agents that can be seamlessly integrated into existing military systems with minimal changes and enable warfighters to work effectively and safely in highly complex and uncertain environments. The agents will utilize a Bayesian network model to learn the probabilistic relationships between cognitive states of a human operator and his/her performance from neural, cognitive and behavioral data streams. Additionally, the Bayesian network model is able to adapt to different military systems based on the profile of users and task environments. In Phase I of this SBIR, we will demonstrate the benefits of configurable agent architecture and interfaces in the Mixed Initiative Experimental (MIX) Testbed a distributed training testbed for human-robot teams. We will further evaluate the feasibility of the Bayesian network modeling for single and multi-modal cognitive state assessment using data collected by UCF-IST.