Adaptive Aiding Using Physiological Operator Functional State Assessment

The future vision of the Air Force depends upon the development and fielding of critical capabilities to sustain and improve its warfighting mission. Key building blocks consist of the capability to harness information to enable better decision making, and the competency to improve standoff and penetration capabilities to reduce vulnerabilities to friendly troops. This area of focus is handled by UCAVs (Unmanned Combat Air Vehicles).

UCAVs, unmanned aerial vehicles with the capabilities to deliver weapons, are being developed by the Air Force to revolutionize tactical airpower by enabling preemptive and reactive suppression of enemy air defense missions. A recent report on a USAF UCAV advanced technology demonstration identified two critical areas that require investigation: dynamic distributed mission/vehicle control, and advanced cognitive aids integration.

The proposed research effort by the hpam Team working for UCAV, examines the physiological impact of workload in executing dynamic supervisory control of distributed vehicles in a simulated UCAV. The research efforts outcome is a proof-of-concept adaptive interface that integrates a predictive physiological model to enable effective performance in the UCAV context. The primary focus is on physiological measures. An inductive approach to associate physiological activity with cognitive task demands using state estimation algorithms and also a deductive approach to model operator physiological states based on dynamical models representing the response of cardiopulmonary and neuromuscular systems to cognitive workload is done.

Project Management:

The UCAV is handled as six separate task objectives:
UCAV environment development
Psychophysiological state estimation algorithm development
Physiological state model development
Adaptive interface design and construction
System integration and evaluation in the FPL
Collaborative teaching of a graduate-level course on adaptive aiding using operator state assessment

UCAV environment development

This objective is focused to be accomplished during the first phase of the project. To meet this objective, we will begin with a detailed hierarchical cognitive task analysis that identifies:

Different mission scenarios that need to be considered to differentiate workload in the context of UCAV control

A normative task-subtask analysis showing what tasks and subtasks need to be performed for a successful conclusion to each scenario, and including an explicit indication of which agents (human and computer) could perform which subtasks

Identification of alternative methods (both performance based and physiological) for detecting the possible occurrence of failures and

Identification of alerting and task re-allocation strategies to deal with predicted problems.

Psychophysiological state estimation algorithm development

Psychophysiological assessment methods measure changes in the operator physiology that are associated with cognitive task demands. To meet this Objective, investigation about the use of psychophysiological state-estimation algorithms toward the design and development of adaptive interfaces in the simulated UCAV environment is done. This effort relies upon inductive state-estimation methodologies, and is highly contingent on task complexity. The research is proposed to be executed in three steps during phase two of the project.

The first step consists of investigating the effectiveness of current algorithms to estimate operator states based on UCAV mission complexity and operator performance.

The second step involves the development of predictive algorithms to anticipate operator needs during the course of UCAV mission execution.

In the third step, indicators within the predictive algorithm that will prompt adaptation in the interface are identified.

Physiological state model development

As a complement to estimation algorithm development objective the second part of the phase two of the project, this objective investigates the use of deductive models to assess operator physiological states. The physiological state model for cognitive workload is categorized into 3 different sections.

Section 1 considers the quantitative analysis of neuromuscular         activity with respect to instrumentation and specific system equations.

Section 2 takes care of the quantitative analysis of cardiopulmonary activity with respect to specific instrumentation systems that provide quantitative data.

Section 3 handles the analysis of cognitive workload with respect to the quantitative physiological reactions.

Adaptive interface design and construction

This objective will be accomplished during phase three of the project. Based on the results of the first phase using analytical and empirical methods to identify methods for detecting and providing adaptive aiding, this objective will require the identification of conceptual approaches for providing aiding when performance and physiological measures suggest a potential problem.

System integration and evaluation in the FPL

While it is expected that development of the adaptive interface and physiological state models would be situated in the context of the UCAV, these elements must still be integrated within the simulation environment. Integration will begin when the adaptive interface, the physiological state model, and the psychophysiological state estimation mechanism has been constructed. This objective is to be accomplished during phase four of the project.

Approximately 500 hours of FPL equipment time is expected to be contributed as in-kind support from AFRL for the project. The US Air Force has used the FPL toward understanding the effects of mental workload using brain electrical activity, heart rate, eye movements, and respiration patterns as indicators of the changing cognitive state of an operator in the laboratory, simulators, and aircraft. The data collected in these settings help define the parameters most useful for evaluating the impact of new or enhanced systems. In the evaluation study, comparison of the two adaptive aiding methods (based on the psychophysiological state estimation algorithm and the physiological state model) to the baseline (e.g., non-aided) interface is done. Quantitative performance variables will be used in the evaluation to compare the systems. This validation test serves as the culmination of this project, and a point of departure for future systems.

Collaborative Teaching

The research conducted through this project will be integrated to teaching during the second year. In the fall quarter of 2002, a course is scheduled to be collaboratively taught between OSU and WSU via the Interactive Video Network (IVN) classroom. The focus of the course will be on the fundamental impact of physiological models on systems design. At WSU, the course will be taught as "HFE 890 Quantitative Methods for Cognitive Modeling". The primary instructor for the course will be Dr. Rothrock. Each investigator on the project will be responsible for at least one week of lecture material.