Non-invasive, Real-Time Measurement of Cognitive Workland for Adaptive Aiding of the Overburdened Decision-Maker

This project will conceptually design and demonstrate the feasibility of a multivariate approach to the real-time measurement of cognitive workload, basedon operator psychophysiology as well as behavior, in the context of task demands. Such as approach will facilitate the development of cooperative human-computer controversy over how to implement such capabilities, including how to ascertain the cognitive status of the human operator and whether and how to adaptively aid the operator "automatically without increasing the risk of catastrophic failure. Part of the difficulty lies in the unreliability of real-time measures of cognitive performance that are based strictly on behavior. The classification system envisioned here will involve the concurrent acquisition and measurement of numerous channels of information, some characterizing the sychophysiological status of the operator, others monitoring the operator's behavioral responses, and still other receiving inputs from the task environment with which the operator is interacting. We will focus initially on characterizing residual capacity during instances of operator overload, although eventually one would hope to cover both ends of the workload continuum. The Phase I wffort will involve an empirical proof-of -concept to demonstrate the effects of workload manipulation on the measures of interest in a low-fidelity simulation. Then we will explore less traditional analytical approaches with measurement approach akin to that which could be eventually implemented in real-time. Finally, we willf ormulate the design of an integrasted data acquisition and analysis workstation that can infer operator cognitive status in real-time. This design, which would be implemented in Phase II, will include a multi-processor hardware configuration, real-time analysis algorithms, and initial decision rules for providing adaptive controil inputs to the system with which the human operator is interacting. The system will be designed to provide all functionality with a common look-and-feel, with an open architecture for future upgrades, and with the ability for use with operators in a variety of task environments.