Glenn F. Wilson, Ph.D, AFRL, Wright-Patterson Air Force Base

 

Glenn F. Wilson, Ph.D, AFRL, Wright-Patterson Air Force Base

Sage 4101

March 19, 2008 11:15 AM - 12:45 PM

The human operator's ability to perform their job can fluctuate from day-to-day and even from moment-to-moment.  Because the cognitive demands of the job can also vary it is possible that the capabilities of the operator are not sufficient to match the job demands.  This can lead to errors when the operator is overwhelmed by the task demands.  Adaptive automation strives to provide automation "on demand" when it is needed by the operator.  Knowing when to present the adaptive aiding is critical.  If it is not presented or presented too late or not when needed it will not assist the operator and may interfere with the task.  Psychophysiological measures, such as heart rate, blink rate and brain activity, can be used to monitor operator cognitive workload.  It has been demonstrated that using psychophysiological measures to initiate adaptive aiding can produce significant improvement to task performance.  Numerous psychophysiological measures have been examined to determine their suitability for use to reveal operator functional state (OFS).  Several classifiers have been tested for their ability to correctly classify OFS and to do this in real time.  Artificial neural networks have been used in our work.  Besides being able to deal with nonlinearities in the psychophysiological data they can provide information about the relative saliency of the input features.  Further, it has been proposed that procedures could be developed that would incorporate this into an overall adaptive aiding system.

 

http://ruccs.rutgers.edu/%7ejacob/Papers/feldman_singh_skeletons.pdf

Add to calendar
Share|