Situational awareness (SA) is a term that has great
intuitive appeal, especially in aviation. Although
often invoked in a descriptive manner, SA is difficult
to precisely define. As a result, a computational,
mathematically defined model of SA has been
developed. The goal of the modeling effort was to
enhance researcher communication and to advance
efforts to improve pilot SA and performance through
improved display design or aircrew procedures.
The Computational Situational Awareness (CSA)
model is composed of two essential features: situational
elements and situation-specific nodes.
Situational elements (SEs) are relevant information in
the environment that define the circumstances (for
example, other aircraft, obstacles, way points, own-ship
parameters). The pilot experiences these elements
through perception, experience, or a preflight
briefing. Each SE has a mathematical weight based
upon its importance in the situation and a
mathematical value based upon one of four levels of
awareness (detection, recognition, identification, and
comprehension). These four levels of awareness
provide a means of quantifying an operator's perception
of the situational elements.
Situation-sensitive nodes are semantically related
collections of SE's. The nodes are defined by the
context of a given task and are weighted by the
overall importance of the node in determining the
level of SA. If the situation changes, then the weights
on the nodes, or the nodes themselves, may change
to reflect accurately the level of SA. SA is the
weighted average of knowledge that the pilot has in
each node, and thus is a measure of the pilot's
perceived SA. The CSA model then subtracts an error
component, based on misidentified SE's or unknown
elements in the environment.
The computational model of situational awareness
was designed to be embedded in an existing
model of human performance, the Man-Machine
Design and Analysis System (MIDAS). MIDAS has
very detailed representations for the two major
components of human-systems integration: (1) the
human operator, and (2) the system, or environment
under study. The human model consists of perception,
and cognitive processes such as working
memory, scheduling, decision making, and long-term
memory. Output measures from the aggregation of
models include task execution time lines, operator
workload, and the aforementioned situation-awareness
construct.
The systems model includes the cockpit, or
workstation, the environment, and the human figure
model. The environmental model consists of elements
in the world with which the simulated operator
or crew station interacts, for example, trees, other
aircraft, tanks, or air traffic control. The human
anthropometric model is Jack®, developed by the
University of Pennsylvania.
The CSA model was developed, integrated into
MIDAS, and tested in simulation. A recent study
evaluated the validity of the CSA when compared
with pilots in a manned simulation with very favorable
results. As shown in figures 1 and 2, both
subjective ratings of SA (Subjective Awareness Rating
Technique: SART) and performance measures of SA
(RMSE altitude) are highly correlated with the predicted
SA levels.
Point of Contact: R. Shively
(650) 604-6249
jshively@mail.arc.nasa.gov
Back To Top
Previous Paper
Return to Revolutionary Technology
Next Paper 
