[ Pobierz całość w formacie PDF ]

through computational analysis. Each element would be represented by an an-
alytical model with which the output could be computed for specified inputs.
The elements would be arranged and interconnected in the form of Fig. 1 and
the total system responses could be computed for specificed inputs. More im-
portantly, the dynamics of the system could be assessed and related to the
characteristics of the elements which form the system. This form of analytical
representation was pioneered by Tustin [38] but has been further developed and
evangelized by McRuer [39] and others. McRuer has had a profound influence
20
on our understanding of the interactive dynamics of the pilotvehicle system,
especially by his dedication to the need for parallel development of analysis and
experimentation. His work stimulated other researchers such as Krendel, [40]
Hall, [41] Anderson, [42] and Neal and Smith [43] to contribute analysis methods
based upon experimental data.
The major difficulty with analysis has been the analytical representation of
the adaptive human pilot. The other elements of the system can be accurately
represented; the pilot and his actions are only partially understood. One trou-
blesome aspect of the pilot is in knowing what variables he is sensing and acting
upon in supplying his corrective control inputs. For example, as he performs
the air-to-air fighter task, he will close different loops during different portions
of the tasks: perhaps normal acceleration and pitch rate for corrective pitch
inputs during the acquisition turn, and pitch angle error and pitch rate during
tracking. What does he sense in landing: altitude, altitude-rate, pitch angle and
rate, ground speed? How do the amounts of each of these vary? What strategy
does he employ to select among the variables as the maneuver progresses, and
what  gains does he use? Work of very high quality has been done in these
areas, but we are still of limited capability in predicting the pilot s dynamic
behavior, especially in new situations. As quoted by McRuer, [39] the words of
Cowley and Skan in their 1930 paper [44] describe the difficulties of pilot-vehicle
analysis even today:  A mathematical investigation of the controlled motion is
rendered almost impossible on account of the adaptability of the pilot. Thus, if
it is found that the pilot operates the controls of a certain machine according to
certain laws, and so obtains the best performance, it cannot be assumed that
the same pilot would apply the same laws to another machine. He would sub-
consciously, if not intentionally, change his methods to suit the new conditions,
and the various laws possible to a pilot are too numerous for a general analysis.
Experimental
Experimental methods are the other means of assessing the quality of the pilot-
airplane combination. Experimentation involves the combination of the pilot
and either the real vehicle or a simulation (ground or in-flight) of the vehicle in
the accomplishment of the real task or a simulation of the real task. For design
purposes, simulation is employed, and the preponderant use is of ground simu-
lation, although in-flight simulation is assuming a growing role during aircraft
development.
There are two general data outputs of the experimental methods: perfor-
mance measurement, and pilot evaluation. Because the pilot is adaptive, per-
formance measurement should include not only how well he is doing (task per-
formance), but also how much effort the pilot is supplying (workload) to achieve
that performance. Workload is used to convey the amount of effort and atten-
tion, both physical and mental, that the pilot must provide to obtain a given
level of performance. [1] The meaningfulness of task performance measurement
data is always bounded by realism of that which is measured. For example,
much effort has been devoted to measuring tracking performance expressed in
21
statistical measures of aim wander taken from 30 to one mim tracking runs on
a nonmaneuvering target; the real fighter pilot deals generally with an aggres-
sively maneuvering target on whom he needs only to achieve the correct solution
long enough to fire. These are really two different tasks. Workload measures
are difficult, too: physical workload of various kinds have been measured; men- [ Pobierz całość w formacie PDF ]

  • zanotowane.pl
  • doc.pisz.pl
  • pdf.pisz.pl
  • moje-waterloo.xlx.pl

    •