Judging from both the explicit and implicit contents of recent postings,
CRM is a very broad topic, and to advance CRM understanding, training,
and practice, there is some need to focus on specific aspects of it. The
purpose of my message is to solicit comments, suggestions, and
assistance on one of those aspects. I realize that this is a departure
from this month's topic, but Neil urged me to post this to the list. [OK
Neil, you asked for it!]
At Oregon State University, we have been studying a process performed on
the flightdeck which we call Agenda Management (AM). AM is a high-level
activity management process which results in the initiation,
interruption, resumption, and termination of multiple, concurrent
activities. It involves managing goals (setting, prioritizing,
cancelling), managing functions or tasks to achieve those goals,
(initiating, prioritizing, terminating), assigning actors (pilots or
automation) to functions, and allocating resources (both human and
machine) to perform them. A more complete description of AM and our AM
research may be found on our Web page at
http://www.engr.orst.edu/~funkk/AM/amsum.htm
(I'll be happy to send an e-mail version to anyone who doesn't have Web
access.)
Of course, Agenda Management is just a name for something that pilots do
all the time, and usually pretty well. AM is an aspect of CRM, or maybe
a collection of aspects. But the concept, suitably formalized, provides
a perspective on a subset of CRM which draws on research in artifical
intelligence and cognitive psychology. As such, it may provide a useful
supplement to the perspectives already represented in the CRM community.
You will have to be the judges of that.
But there is already some evidence that AM may be a useful concept. We
have reviewed hundreds of aircraft accident and incident reports and
found many cases where tasks were started too soon or too late, were
misprioritized, or were terminated prematurely or too late. Our
tentative conclusion is that AM is a significant factor in flight
safety.
Our long term goal is to develop means of facilitating AM. For example,
we are now working on an experimental computational aid with funding
from the NASA Ames Research Center.
But the greatest need is to understand AM better, perhaps in its
relation to CRM as a whole. To further our understanding, we hope to
perform detailed analyses of flightdeck data. CVR transcripts and flight
data recorder data from accident reports provide us with the kind and
detail of data needed to construct activity timelines. But analysis of
accident report data may only tell us about how NOT to do AM. We want to
understand how AM SHOULD be done. So we are looking for flight simulator
data. At minimum, this data should contain voice recordings or
transcripts and some aircraft state data. Video tapes would be helpful.
We need to be able to determine who (or what) was doing what when.
I think that LOFT tapes would do the trick, but LOFT tapes are
confidential. Even tapes from simulator studies at NASA can't leave NASA
(though we're working on that). We've seen flight certification data,
but we don't have access to much of that (and it may be biased anyway).
Do any of you have data that might be useful to us? Do you have CRM
training tapes used to illustrate good/bad CRM that would have enough
detail and that you would be willing to share with us? Do you have
suggestions on who to contact and what to ask for? Would any of you with
the resources (simulators, personnel, etc.) be interested in
collaborating on a new study to develop the data we need?
Thanks for your attention. We're definitely in the learning mode and
will appreciate any help we can get.
Ken
-- Kenneth H. Funk II Industrial & Manufacturing Engineering Oregon State University Corvallis, Oregon, USA funkk_at_engr.orst.edu http://www.engr.orst.edu/~funkk