Evaluation of a Precision Hover Task Using Time-Varying Cutoff Frequency

skillfulbuyerUrban and Civil

Nov 16, 2013 (3 years and 10 months ago)

72 views

Evaluation of a Precision Hover Task
Using Time
-
Varying Cutoff Frequency



Amanda K. Lampton, Ph.D.

David H. Klyde

Daniel J. Alvarez

P. Chase Schulze

Peter M. Thompson, Ph.D.

Chi
-
Ying Liang, Ph.D.


Systems Technology, Inc.

Hawthorne, CA







Presented to

67th Annual Forum of the American Helicopter Society

Virginia Beach, VA

May 4 2011

4 May 2011

67th Annual Forum of the American Helicopter Society

ACKNOWLEDGEMENTS


NASA Research Announcement under Topic A.3.6, “Rotorcraft
Flight Dynamics and Control”


NASA
:

William Decker (COTR) and VMS Team


HOH AERONAUTICS
:

Dave Mitchell (PI) and Tom Nicoll


AMRDEC
:

Jeff Lusardi


ADVANCED ROTORCRAFT TECHNOLOGIES

2

PRESENTATION OUTLINE


Introduction


Baseline & Added Dynamics


Time
-
varying Cutoff & Power Frequency


Piloted Simulation Description


Precision Hover Task Analysis


Conclusions


4 May 2011

67th Annual Forum of the American Helicopter Society

3

INTRODUCTION


The objective of this research program was to
investigate pilot thresholds of detection as higher
-
order
dynamics were introduced to a baseline rotorcraft model


The investigation took the form of a piloted simulation
that was conducted in February 2009 using the NASA
Ames Research Center Vertical Motion Simulator (VMS)


The focus of this paper is to assess the utility of time
-
varying cutoff and power frequency as a means of
differentiating run
-
to
-
run and pilot
-
to
-
pilot differences in
pilot
-
vehicle system behavior


4 May 2011

67th Annual Forum of the American Helicopter Society

4

BASELINE ROTORCRAFT
DYNAMICS


The baseline helicopter model used was a modified OH
-
6A
with the cross coupling and higher order terms of the model
removed.


Removing the cross coupling terms helped to insure that any
changes seen in the dynamics of the helicopter by the pilot
were due to the added dynamics only and not from potential
cross
-
coupling.


The baseline rotorcraft dynamics represented uncoupled pitch
and roll dynamics resulting in first order pitch and roll rate
command systems.


The pitch and roll damping derivatives were set to provide one
high and one low bandwidth configuration in each axis.


The simplified model was valid only for low speed and hover.

4 May 2011

67th Annual Forum of the American Helicopter Society

5

ROTORCRAFT ADDED
DYNAMICS


The added dynamics took the form of a second order
lead/lag or lag/lead transfer function filter.


The filter was cascaded with the baseline rotorcraft
dynamic model.


The gain, pole frequency, and zero frequency could all
be varied individually.


The pole and numerator damping were set to the same
initial value and were then varied together.


The added dynamics had the following form:


4 May 2011

67th Annual Forum of the American Helicopter Society

6

,
,
z
z n
p np
K
 
 
 
 
 
 
NASA AMES VERTICAL MOTION
SIMULATOR (VMS) FACILITY


The VMS features significant vertical and
horizontal motion that makes the facility
ideal for conducting low speed rotorcraft
evaluation tasks.


The simulator cab was configured with
standard rotorcraft controls (center cyclic,
collective, and pedals) and displays for
forward, side, and chin bubble windows

4 May 2011

67th Annual Forum of the American Helicopter Society

7

NASA Photo & Illustration

PRECISION HOVER TASK


The evaluation task used was the precision hover task as
specified in ADS
-
33E
-
PRF.


The objective of the task is to evaluate the ability of the rotorcraft
to transition from translating flight to a stable hover over a
designated point, accurately and with adequate aggressiveness.


4 May 2011

67th Annual Forum of the American Helicopter Society

8

NASA Photo

EVALUATION PROCEDURES


The pilots were familiarized with the evaluation task
using the baseline (no added dynamics) configuration.


Some adjusting of the lateral and longitudinal cyclic
gains was made, if necessary, to ensure that the pilot
felt comfortable with the baseline configuration such that
the task performance was considered “good.”


The baseline configuration was often repeated to
recalibrate the pilot to known “good” dynamics.


For an individual evaluation case, the pilot was given a
unique zero/pole combination with a shared damping
ratio.

4 May 2011

67th Annual Forum of the American Helicopter Society

9

POST RUN PILOT QUESTIONS


At the conclusion of the task, the pilot was given the
following three questions:


Did you notice the added dynamics?


If you noticed the added dynamics, did they affect the task?


If the added dynamics affected the task, did they improve or degrade task performance?


The damping ratio would then be changed based on
the responses of the pilot to the above questions.


This process would be repeated until a more or less
complete set of responses were achieved for that
unique zero/pole pair.


A new zero/pole pair would then be introduced and
the process would begin again, usually with the
baseline configuration.


4 May 2011

67th Annual Forum of the American Helicopter Society

10

PILOTED SIMULATION SUMMARY


The piloted simulation was conducted over a three week
period in February 2009.


Five rotorcraft test pilots participated in the program
resulting in over 1100 evaluation runs.


These runs consisted of many combinations of added
dynamics in the pitch and roll axes with the added
dynamics being introduced as both time
-
invariant and
time
-
varying cases.


While the majority of runs were conducted with full VMS
motion, a number of runs were conducted with no
motion.


The full motion of the VMS was required, particularly for
the higher frequency added dynamics cases.


4 May 2011

67th Annual Forum of the American Helicopter Society

11

CLASSIC CUTOFF FREQUENCY


A spectral analysis method for determining the pilot
operating frequency for pilot
-
in
-
the
-
loop flying tasks


An alternative measure when it is not possible to
determine pilot
-
vehicle crossover frequency directly


A quantitative measure of pilot stick activity derived
by examining controller input power versus
frequency


Defined as the frequency at which the integral of the
power spectral density (PSD) is half its total value


4 May 2011

67th Annual Forum of the American Helicopter Society

12

TIME
-
VARYING CUTOFF
FREQUENCY


Wavelets provide a means of calculating the time
-
varying power or auto spectrum, called a scalogram


Rather than using the auto spectrum averaged over the
entire run, the time
-
varying cutoff frequency is calculated
by numerically integrating the power over the frequency
range for each time increment of the scalogram

4 May 2011

67th Annual Forum of the American Helicopter Society

13











1
2
0
1
2
0
1
2
1
2
0.5
total
t
G d
G d
t
t
t
t














POWER FREQUENCY


A limitation of the time
-
varying cutoff frequency is the
lack of a relationship to the magnitude of the power
spectrum


This can result in a range of behavior that does not fully
correlate to the time
-
varying behavior seen in the
associated scalogram


The power frequency metric marries the two by
multiplying the cutoff frequency by the peak magnitude
at each time slice


4 May 2011

67th Annual Forum of the American Helicopter Society

14







max
1000
cutoff
G
t G t
t





ANALYSIS: SELECTED

ADDED DYNAMICS CASES

Pilot

Configuration

Added Dynamics
Damping

A

Gain (0.56)

Lead (1.5 r/s)/Lag (2 r/s)

0.5 → 0

C

Lead (4 r/s)/Lag (4.5 r/s)

0.3 → 0

D

Gain (1 → 0.79)

Lead (4 r/s)/Lag (4.5 r/s)

0.4 → 0.01

4 May 2011

67th Annual Forum of the American Helicopter Society

15

ANALYSIS:

Pilot A Time
Histories

4 May 2011

67th Annual Forum of the American Helicopter Society

16

Lateral
Cyclic

Roll Rate

Roll Rate

Lateral
Cyclic



㴠呖

Baseline

ANALYSIS: Pilot A Baseline

4 May 2011

67th Annual Forum of the American Helicopter Society

17

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS: Pilot A Time
-
Varying

4 May 2011

67th Annual Forum of the American Helicopter Society

18

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS:

Pilot C Time
Histories

4 May 2011

67th Annual Forum of the American Helicopter Society

19

Lateral
Cyclic

Roll Rate

Roll Rate

Lateral
Cyclic



㴠呖

Baseline

ANALYSIS: Pilot C Baseline

4 May 2011

67th Annual Forum of the American Helicopter Society

20

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS: Pilot C Time
-
Varying

4 May 2011

67th Annual Forum of the American Helicopter Society

21

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS:

Pilot D Time
Histories

4 May 2011

67th Annual Forum of the American Helicopter Society

22

Lateral
Cyclic

Roll Rate

Roll Rate

Lateral
Cyclic



㴠呖

Baseline

ANALYSIS: Pilot D Baseline

4 May 2011

67th Annual Forum of the American Helicopter Society

23

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS: Pilot D Time
-
Varying

4 May 2011

67th Annual Forum of the American Helicopter Society

24

Lateral
Cyclic

Roll Rate

Cutoff Frequency

Scalogram

Power Frequency

ANALYSIS SUMMARY

4 May 2011

67th Annual Forum of the American Helicopter Society

25

Cutoff Frequency

Power Frequency

CONCLUSIONS


Time
-
varying power frequency provides a promising new
means to differentiate run
-
to
-
run and pilot
-
to
-
pilot
characteristics in pilot
-
vehicle system behavior for closed
-
loop
tasks.


The dependence of the time
-
varying power frequency on the
time, magnitude of the power spectra density, and the
distribution of the power over the frequency range clearly
show the effects of the change in dynamics and the
transitions between phases of the task.


Time
-
varying cutoff frequency is a less clear measure of pilot
effort. Without the dependence on the magnitude of the power
in the scalograms, it is difficult to judge pilot effort and to
extract clear differences between runs and pilots.


4 May 2011

67th Annual Forum of the American Helicopter Society

26