United States Department of Transportation
i
Hidden Markov Models as a tool to measure pilot attention switching during simulated ILS approaches
-
2003-04-14
-
[PDF-186.84 KB]
Details:
-
Creators:
-
Corporate Creators:
-
Subject/TRT Terms:
-
Resource Type:
-
Corporate Publisher:
-
Abstract:The pilot's instrument scanning data contain information about not only the pilot's eye movements, but also the pilot's
cognitive process during flight. However, it is often difficult to interpret the scanning data at the cognitive level
because: 1) some instruments provide partially-redundant information, and 2) some instruments display more than one
type of information. To avoid these problems, a new modeling and analysis technique is demonstrated that looks at the
scanning data from the task attention-level as well as from the instrument-fixation level. Basic principles of attitude
instrument flying were incorporated to construct a task-instrument framework, and a Hidden Markov Model (HMM)
was used to estimate hidden task transitions. An ILS simulation experiment was conducted, and the task sequence
estimated by the HMM matched 79-92% of the task epochs verbally reported by the pilot. The HMM and traditional
instrument fixation analyses complement each other, and successfully detected effects of display format changes that
did not produce significant changes in flight technical error and subjective workload. Potential advantages of this
method for the analyses of advanced cockpit displays (e.g., primary flight displays, head-up displays) are discussed.
-
Format:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
