670 HELICOPTER NIGHT PILOTAGE

HELICOPTER NIGHT PILOTAGE

In 1971, the United States Army determined that, in order to
survive on the modern battlefield, tactical helicopters had to
fly very near the ground and hide behind terrain contour or
trees. Flying at very low altitude, masked by hills and trees,
was required in order to overcome the threat of enemy ground
to air weapons.
Flight to and from the battle area is at high speed and
constant altitude above the ground, generally less than thirty
feet above the terrain or local obstacles. This is called contour
flight. Flight in the battle area is nap-of-the-earth (NOE).
During NOE flight, at least part of the aircraft is below tree-
top level, and the aircraft flies around obstacles rather than
over them in order to remain hidden. NOE and contour flight
requires night imaging sensors with field of view (FOV) and
resolution sufficient to allow the pilot to fly the aircraft near
trees and other ground obstacles.
The night pilotage task is very demanding on both the avi-
ator and the helicopter night sensors. A helicopter night pilot-
age sensor should allow the pilot to fly ‘‘heads up and eyes
out’’; the system should provide the same type of contextual
information at night which allows the pilot to orient and fly
the aircraft during the day with unaided vision. The sensor
should provide an image that permits the pilot to perform
precision aircraft movements in a confident and aggressive
manner. The sensor should permit the pilot to discern terrain
features for navigation, select low-level flight paths, and de-
tect possible threats. A good pilotage sensor will also max-
imize the fraction of time that at least minimal performance
can be gained from the sensor in order to execute a mission.

NIGHT PILOTAGE SENSORS CURRENTLY IN USE

Image Intensifiers
The first fielded imaging aid used for low-level night pilotage
was the AN/PVS-5 Night Vision Goggle which was adopted
from ground use. The AN/PVS-5 goggle is shown in Fig. 1.
This sensor uses image intensifier (I 2) tubes which amplify
moonlight and starlight. The goggle amplifies visible light
and provides a considerably brighter image to the pilot than
would be available without the goggle.
The goggle provides a binocular image (an image to both
eyes) with 40⬚ circular FOV. To illustrate this field of view, a
19-inch television set viewed from 21 inches would provide
about the same field of view to the eye as the goggles. The
goggle image, however, is optically projected as a virtual im-
age that appears to be outside the aircraft; this relieves eye
strain and makes the image appear more natural. The image
is unity magnification, meaning that objects appear life-sized.
Under optimal light conditions, the AN/PVS-5 goggles
have a limiting resolution of 0.7 cycles per milliradian (cy/

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.
 HELICOPTER NIGHT PILOTAGE 671

Figure 1. The AN/PVS-5 goggle provides a good image with moon- Figure 2. The ANVIS goggle provides a good image with moonlight
light illumination. In use, it covers the entire upper portion of the or starlight illumination. The pilot can view instruments by looking
face. under the goggle.

mrad) which is equivalent to a visual acuity of about 20/50. gyro horizon without looking inside at the cockpit instru-
(When an optometrist says that you have ‘‘20/50 vision,’’ he ments. Figure 3 illustrates symbology superimposed on AN-
means that you can read the same size letters at 20 feet as VIS imagery. The HUD allows the pilot to keep ‘‘heads up and
are legible to most people at 50 feet. The human eye resolu- eyes out,’’ because the pilot need not focus his eyes and atten-
tion at the 20/20 level corresponds to the ability to resolve tion inside the cockpit to view important instrument infor-
roughly one minute of arc.) mation.
Experience with the ground goggle showed it to be a sig- The primary problem with using ANVIS on helicopters is
nificant aid for night flight. Two significant problems were lack of compatibility with the cockpit instrument lighting.
encountered, however. In use, the ground goggle covers the Modern image intensifiers amplify ambient light 2000 to 3000
entire upper portion of the face, so that the pilot viewed both times; cockpit lights can blind the goggles due to reflected
the outside world and aircraft instruments through the gog- glare off the canopy or off other objects in the cockpit. The
gle. The goggle optics could not be focused to simultaneously problem is corrected by adding a spectral filter to ANVIS
show both the nearby instruments and the outside world. The which rejects blue-green light, and only blue-green instru-
second problem with the ground goggle was that it provides a
good image only when the moon is up; flying with these gog-
gles was difficult under starlight illumination conditions.
The development of an I 2 goggle specifically designed for
aviation use was initiated in the late 1970s. The new goggle
was designated the AN/AVS-6 Aviator’s Night Vision System
(ANVIS). ANVIS mounts to the pilot’s helmet as shown in
Fig. 2 and allows the pilot to view his instruments by looking
under the goggle. ANVIS can also be rotated up to a stow
position on top of the helmet, leaving the pilot’s vision com-
pletely unobstructed.
ANVIS provides a good image under starlight illumination
conditions. In addition to being more sensitive than the AN/
PVS-5 in responding to visible light, the ANVIS spectral band
encompasses more of the ambient light available at night.
ANVIS responds to near infrared light as well as to visible
light. ANVIS provides a 40⬚, binocular, unity magnification
image with better resolution than the original ground goggle.
Under optimal illumination conditions, ANVIS limiting reso-
lution is about 0.9 cy/mrad corresponding to a limiting acuity
of 20/40.
The AN/AVS-7 Heads Up Display (HUD) was added to
ANVIS in the early 1990s; it is a small apparatus which
clamps onto one of the ANVIS oculars. The HUD superim- Figure 3. Flight symbology is superimposed on the ANVIS imagery;
poses instrument symbology on goggle imagery, allowing the the pilot does not need to look inside the cockpit to see important
pilot to see important information like altitude, heading, and aircraft status information.
672 HELICOPTER NIGHT PILOTAGE

ment lighting is used on the newer Army helicopters. Red

light is avoided because ANVIS is quite sensitive to red light.
Lighting requirements for ANVIS compatibility are discussed
in ref. 1.

Thermal Imagers
In 1973, development was initiated on the first thermal im-
ager for pilotage use. The AN/AAQ-11 Pilot’s Night Vision
System (PNVS) was developed for the AH-64 Apache Ad-
vanced Attack helicopter. PNVS is a gimbaled thermal im-
ager mounted on the nose of the helicopter. The position of
the PNVS on the helicopter is shown in Fig. 4. The PNVS
images 8 애m to 12 애m thermal energy (that is, heat) and
provides a 40⬚ horizontal by 30⬚ vertical FOV.
The pilot is in the cockpit, while the PNVS thermal imager
is on the nose of the aircraft. The system hardware must pro-
vide some means of pointing the sensor where the pilot wants
to look and some means to remote the thermal image back
to the pilot in the cockpit. Figure 5 illustrates how this is Figure 5. Pilot wears a helmet mounted display in front of right eye;
accomplished on Apache. he uses this to view the PNVS thermal imagery. A helmet tracker
A helmet tracker slaves the sensor line of sight to the pi- turns the PNVS sensor to match the pilots head movement.
lot’s head. The pilot wears a helmet-mounted display through
which he views the thermal image. The helmet display pro-
jects a virtual image which appears to be outside the aircraft. used by the copilot/gunner to locate and engage targets. How-
The helmet-mounted display is monocular, viewed with the ever, the TADS thermal imager has three fields of view with
right eye only, and provides the same 30⬚ vertical by 40⬚ hori- the wide field of view identical to the PNVS field of view. The
zontal field of view as the sensor. The system therefore pro- copilot/gunner can use the TADS image in a pilotage mode in
vides a unity magnification, thermal image of the world which exactly the same way that the pilot uses the PNVS. A helmet
the pilot can orient by moving his head. tracker senses the copilot’s head motion and moves the TADS
A second thermal imager is available on the Apache heli- to align the line of sight of the thermal imager. The copilot
copter. The second thermal imager is one of several sensors views the image via a helmet-mounted display.
in the AN/ASQ-7 Target Acquisition and Designation System Heads-up instrument symbology is an integral part of the
(TADS); the TADS is the large, barrel shaped object located PNVS and TADS systems on the Apache helicopter. Both pi-
below the PNVS shown in Fig. 4. This imager is normally lot and copilot can view important flight and status informa-
tion superimposed on the thermal imagery. With symbology
superimposed on his night vision imagery, the pilot does not
have to focus his eyes inside the cockpit to determine critical
information such as altitude, heading, or caution status.

Combinations of Thermal Imagers and Image Intensifiers

In 1987, an adapter was designed to permit the ANVIS to be
mounted on the Apache copilot’s helmet. The adapter allows
the ANVIS to be mounted simultaneously with the Apache
helmet display, although ANVIS and the helmet display can-
not be viewed simultaneously. When the copilot is using AN-
VIS, the TADS thermal imagery and symbology can be viewed
on a panel display by looking under the ANVIS. The copilot
can use the ANVIS imagery and periodically cross reference
the thermal imagery as a safety check. If the copilot is using
the helmet-mounted display and TADS thermal sensor, the
ANVIS is placed in the stow position on top of the helmet.
In the late 1980s, the Helicopter Night Vision System
(HNVS), AN/AAQ-16, was fielded on some UH-60 Blackhawk
Utility helicopters and on some CH-47 Chinook Cargo helicop-
ters. The HNVS is a thermal imager which operates on simi-
lar principles to the PNVS and the TADS. The HNVS is
mounted on the nose of the aircraft and is viewed via a panel-
mounted display in the cockpit. The HNVS is not head
Figure 4. The PNVS thermal imager mounted on the front of the tracked, but can be pointed by a hand controller. The sensor
Apache Helicopter. The TADS system is the barrel-shaped object with has two fields of view. The wide FOV is 30⬚ vertical by 40⬚
two windows mounted beneath the PNVS. horizontal; the narrow FOV is 5⬚ vertical by 7⬚ horizontal.
 HELICOPTER NIGHT PILOTAGE 673

Both pilot and copilot use ANVIS to fly. The panel dis- of fiberoptic bundles with the core etched away. The plate has
played HNVS imagery is used to cross reference and verify millions of channels (holes) with photoemissive material on
the information provided by the ANVIS. The aviators use the inside of the channels. Each face of the MCP is metalized,
HNVS as a backup, and as a cross reference for terrain avoid- and a high voltage is applied across the plate. As electrons
ance, target location, check point verification, and during low strike the inside of the MCP channels, secondary electrons
illumination or poor visibility conditions where ANVIS vision are emitted. Multiple secondary electrons are emitted for each
is degraded. cathode electron. The secondary electrons are accelerated by
The newest Army helicopter, currently in development, is the voltage along the channel, the secondary electrons strike
the RAH-66 Comanche; Comanche is a reconnaissance and the channel wall and cause more electrons to be emitted, and
light attack helicopter. The Comanche Night Vision Pilotage the electron multiplication process is repeated.
System will integrate an advanced, high -resolution thermal The amplified electrons from the MCP are accelerated to
imager, an I 2 camera, and flight symbology into a single pack- the phosphor, where a brighter version of the cathode image
age. The pilotage sensors will be mounted on the nose of the is formed. The fiberoptic twist erects this image. The eyepiece
aircraft in a manner similar to Apache; however, the nose magnifies the image for presentation to the eye. ANVIS pro-
turret will include both thermal and I 2 sensors. The pilot will vides a scene to eye light gain of about 3000. In the absence of
wear a binocular helmet display rather than the monocular fog or obscurants, ANVIS performs well under clear starlight
display worn by Apache aviators. The field of view of the illumination. Generally, ANVIS provides good imagery with
NVPS with the new helmet-mounted display will be 30⬚ verti- naked-eye visibility exceeding 200 m to 300 m and minimum
cal by 52⬚ horizontal. light levels of 7E-5 footcandles (2).

Thermal Imagers
SENSOR THEORY OF OPERATION
Thermal imagers like the Apache helicopter PNVS detect ra-
Image Intensifiers diation in the 8 애m to 12 애m spectral band. This band is
chosen because the atmosphere has a ‘‘window’’ where the
The image intensifiers used in ANVIS amplify ambient light, transmission of thermal energy is good. Everything near room
moonlight, and starlight, at spectral wavelengths between 0.5 temperature radiates at these wavelengths. The emissivity of
and 0.9 애m. A schematic of a goggle ocular is shown in Fig. natural objects is generally above 70%; most human-made ob-
6; binocular goggles use two oculars to provide an image to jects are also highly emissive. It should be noted, however,
both eyes. An inverted image of the scene is formed on the that thermal sensors derive their images from small varia-
cathode by the objective lens. The cathode emits photo elec- tions in temperature and emissivity within the scene. Typi-
trons; the shot noise associated with cathode photoelectrons cally, the thermal scene is very low contrast even under good
dominates the performance of image intensifiers. Photoelec- thermal viewing conditions. Scene thermal contrast is af-
trons from the cathode are accelerated to the microchannel fected by the amount of solar heating during the day. Ther-
plate (MCP) by a voltage difference applied between the cath- mal contrast is decreased by the presence of clouds. Thermal
ode and MCP. contrast can be poor at night, particularly after extended peri-
The MCP acts as an electron multiplier and provides most ods of clouds or precipitation.
of the gain of the I 2 tube. A detail of the MCP is shown at the In current thermal imagers like the PNVS, a linear array
bottom of the Fig. 6. The MCP is a thin, glass plate made up of infrared detectors is used. Figure 7 illustrates the theory

Objective
lens Cathode MCP Phosphor Fiber optic
twist Eyepiece

Electrons

Microchannel plate
(MCP)

Figure 6. Theory of operation for an image

intensifier. The microchannel plate is illus-
trated at the bottom.
674 HELICOPTER NIGHT PILOTAGE

Scan
mirror

Afocal Detector
optics Imaging array
lens

Light
(thermal energy)

Electric reformat
Figure 7. Theory of operation for a thermal imager. and display

of operation. The afocal optics provide a magnified image of sensor to deliver the desired visual information; these trades
the scene at the scan mirror. The linear array of detectors is do not relate to the ability of the entire weapon system to
scanned over the image by the oscillating mirror. The image accomplish a mission.
is formed by rapidly sampling each element of the detector When there is good thermal contrast in the scene, and in
array as it is scanned over the whole image area. A video the absence of fog, heavy rain, or snow squalls, the PNVS
image is formed electronically from the detector samples; the thermal imager supports terrain (NOE and contour) flight.
video image is viewed via the helmet-mounted display. Good thermal contrast occurs when there has been clear
The linear array in PNVS has 180 detectors; interlace is weather with sunshine for at least a few hours during the
used to generate 360 active lines in the image. Interlace is day, heating up objects in the background scene. If there has
achieved by nodding the scan mirror alternately up and down been no sunshine during the day, or if there has been only a
a small amount after each sweep of the field of view. little sunshine followed by heavy rain or hours of drizzle, the
Detector noise dominates the performance of these im- thermal contrast will be poor, leading to poor visual flying
agers. PNVS provides usable imagery with tree to ground conditions.
equivalent blackbody temperature differences greater than Further, the thermal radiation which PNVS images is at-
0.3 K; performance with less than 0.1 K temperature differ- tenuated by heavy fog and by the atmospheric water vapor
ence is poor (2). content found with heavy rain and persistent drizzle. Image
contrast might be poor even when the scene is providing a
usable thermal signature. Thus, poor local weather, such as
PILOTAGE SENSOR PERFORMANCE
patches of fog or squalls, may make terrain flight difficult at
the midpoint of a flight, even though conditions are good at
The performance of a pilotage aid depends on the image deliv-
the starting point and destination.
ered to the pilot during flight. Depending on the weather and
ANVIS performs well under clear starlight conditions but
other factors, the image can fade, become very noisy, and even
becomes marginal to unusable under overcast starlight condi-
disappear completely. The image quality of image intensifiers
and thermal imagers is affected by ambient atmospheric con- tions. Heavy fog will shut down ANVIS. Even a moderate fog
ditions and the nature of the local environment. The I 2 image can severely degrade imagery if flight is toward the moon;
quality depends on available illumination from the moon and scattered light from the fog can severely degrade contrast and
stars, on atmospheric visibility conditions, and on the diver- mask the view of the terrain. Also, ANVIS tends to ‘‘bleach
sity and contrast of ground objects in the local area. Thermal out’’ or shut down when bright lights are in the field of view;
image quality depends on thermal contrast within the scene this occurs around city lights, when flying toward the moon if
and on atmospheric transmission in the thermal spectral it is low on the horizon, and under dawn and dusk conditions.
band. Thermal contrast is increased by solar heating during Flying over land that has no features, such as the sand
the day and is reduced by heavy or prolonged cloud cover or dunes of Saudi Arabia, presents a challenge; judging distance
precipitation. and closure to the ground requires scene detail. Areas devoid
User surveys were conducted in 1987, 1990, and after De- of distinguishable features, such as snow fields, lakes, and
sert Storm in 1992 (3–6). Structured flight evaluations have dry lake beds, will provide poor imagery for terrain flight. Un-
also been performed (2,3,4,7). These surveys and evaluations der these circumstances, the availability of flight symbology
provide insight into the environmental conditions under is critical.
which the pilotage systems perform well. Pilots express strong feelings that thermal sensors and im-
While it is straightforward to define good and poor weather age intensifiers are complimentary and that both are needed
and environment conditions for ANVIS and PNVS usage, it is for night contour and NOE flight. The combination supports
very difficult to define the conditions which are safe. An avia- flight under a wider range of conditions than either alone,
tor will change the aircraft airspeed, altitude, and flight pro- although environments certainly exist where even the combi-
file as needed to adapt to the conditions encountered. As night nation will not support terrain flight.
sensor imagery degrades, the pilot will also depend more on Also, each sensor brings a unique capability to the aircraft.
the instruments and the HUD symbology. The engineering The two sensors operate in different spectral bands and de-
trades for a night vision sensor relate to the ability of the pend on different physical principles for performance. The
 HELICOPTER NIGHT PILOTAGE 675

ability of the aircrew to detect wires and other obstacles is Table 1. 1987 Survey: Pilot Rating of PNVS and ANVIS FOV
significantly enhanced. Even on poor thermal nights, the and Resolution
PNVS and HNVS provide a good capability to perceive and Sensor/Feature Good Adequate Inadequate
react to close in targets. Even on nights with poor illumina-
PNVS FOV 5 35 9
tion, ANVIS gives the ability to see town lights and therefore
PNVS Resol. 1 18 30
provides navigational aid; because ANVIS can see aircraft ANVIS FOV 9 17 3
running lights, it also provides a better ability to fly formation ANVIS Resol. 13 13 3
as well as safety from collision with other aircraft.

DATA RELATING TO DESIGN IMPROVEMENTS

before expanding FOV. The pilots are interested in increased
PNVS FOV but only in combination with improved image
On the basis of feedback from pilot interviews, current night
quality.
vision sensors like the ANVIS, PNVS, TADS, and HNVS pro-
A summary of the responses to each survey is given below.
vide a significant improvement in mission effectiveness over
The 1987 survey queried 49 Apache helicopter pilots, all
previous techniques of flying and fighting at night. Apache
with PNVS thermal imager experience; 29 of these aviators
aviators stated that the thermal pilotage and targeting sen-
had ANVIS experience (3,4). When given an open choice of
sors on Apache (the PNVS and TADS systems) completely
which sensor they preferred, 42 of 49 wanted both PNVS
changed their capability to fight at night so that comparisons
and ANVIS.
to previous aircraft were not meaningful. It is also clear from
The Apache crews were asked to give an overall rating for
the pilot surveys, however, that further enhancement of night
PNVS and ANVIS as to adequacy of FOV and resolution (im-
effectiveness can be gained from further hardware devel-
age quality); they were to answer based on their total flight
opment.
experience. Table 1 summarizes how many pilots rated FOV
In recent years, the quality of both image intensified and
and resolution as good, adequate, and inadequate. In general,
thermal imagery has improved substantially. Even with ad-
the pilots rated the PNVS FOV as adequate but the resolution
vanced technology, however, optimizing the design of elec-
as inadequate. They rated both the FOV and resolution of
trooptical pilotage sensors involves trade-off of resolution,
ANVIS as adequate.
field of view, and sensitivity. At any given level of technology,
The large majority of Apache aviators, 45 out of 49, would
for example, an increase in the sensor field of view requires
improve PNVS resolution before expanding FOV. The opinion
a decrease in sensor resolution or a decrease in sensitivity
on ANVIS was about evenly split between improving resolu-
or both.
tion and FOV. However, two cautions were emphasized by the
Further, the optimum performance trade-off of imaging
respondees. First, these numbers do not reflect a lack of inter-
sensor parameters depends on specifying the visual task.
est in increased FOV if it accompanies improved image qual-
Night helicopter pilotage involves many visual tasks. Flying a
ity. Second, the user will not accept a smaller FOV than cur-
helicopter near the ground involves judging distance, judging
rently provided.
closure to terrain or terrain objects, maintaining orientation
The 1990 survey involved 52 ANVIS aviators from three
of the aircraft, looking for a suitable flight path, searching for
units flying a variety of missions (5). Twenty of the ANVIS
obstacles and threats, and other visual tasks.
aviators regularly used the HNVS thermal imager in addition
Over the years since the mid-1970s, responsible Army or-
to ANVIS. Twenty-one PNVS aviators were also surveyed;
ganizations have undertaken field surveys of operational us-
eighteen of the PNVS aviators also used ANVIS. Again, when
ers, flight evaluations, and flight experiments in order to de-
given an open choice of sensor, the overwhelming majority
velop design criteria for helicopter night pilotage systems.
chose a pilotage system with both thermal and image-intensi-
These efforts have focused on determining the fraction of time
fied imagery.
that existing pilotage sensors support mission accomplish-
The aviators were asked to give an overall rating for PNVS
ment and on finding sensor design parameters which optimize
and ANVIS as to adequacy of FOV and resolution (image
flight handling. These efforts are summarized.
quality); they were to answer based on their total flight expe-
rience. Table 2 below summarizes their answers.
User Feedback on FOV and Resolution
Seventeen of the twenty-one Apache aviators would im-
In each of the three surveys taken between 1987 and 1992, prove PNVS resolution rather than expanding FOV with the
the aviators were asked to answer questions, based on their current resolution. Fifty of the ANVIS aviators would expand
total flight experience, about needed design improvements in ANVIS FOV if the current ANVIS resolution could be main-
field of view and resolution for ANVIS and PNVS. In an oper- tained.
ational context, sensor resolution refers to image quality and
therefore depends on the sensor sensitivity as well as the opti-
cal resolving power of the sensor. Table 2. 1990 Survey: Pilot Rating of PNVS and ANVIS FOV
The results of all the surveys are consistent and can be and Resolution
summarized as follows. Based on total flight experience, pilots
Sensor/Feature Good Adequate Inadequate
rate both the FOV and the resolution of ANVIS as acceptable.
Pilots would choose to expand ANVIS FOV but not at the ex- ANVIS FOV 16 45 8
pense of current image quality. On the basis of total flight ANVIS Resol. 32 36 1
experience, pilots rated the PNVS FOV as adequate but the PNVS FOV 2 18 1
PNVS Resol. 0 9 10
resolution as inadequate; they would improve image quality
676 HELICOPTER NIGHT PILOTAGE

The 1992 survey was conducted after Desert Storm (6). No During 1985, a flight experiment was conducted by the
area is as devoid of distinguishable terrain features on such NASA Ames Research Center to determine the visual cues
a scale as Saudi Arabia. The sand dunes lacked almost any essential for low speed and hover flight (9). This test was con-
vegetation and had rises and falls varying as much as 75 feet. ducted in order to determine the importance of field of view
The lack of features made the terrain relief imperceptible and resolution on the fidelity of flight simulators. The vari-
through the night vision sensors. This was a difficult area in ables in this flight test were field of view, the amount of mac-
which to use night vision sensors. rotexture (large objects), and the amount of microtexture (fine
Of 66 aviators surveyed, 70% judged ANVIS performance detail) in the imagery. Field of view was varied by masking
in Saudi Arabia to be good or adequate. What should be noted portions of the windscreen. Microtexture was varied with a
is that the 30% inadequate rating was never experienced else- set of liquid crystal goggles which selectively fogged the im-
where. Of the 34 Apache aviators surveyed, 70% rated the age. Macrotexture was varied by changing flight location and
PNVS performance in Saudi Arabia as good or adequate. by laying objects like tires on the ground near the flight path.
Thermal conditions were better at the beginning of the war, The test fields of view ranged from a 10 by 14⬚ rectangular
and image intensifier conditions were better at the end of the window to a multiwindowed case encompassing 9000 square
war. Aviators with a choice used both systems about half the degrees. Two resolutions were used: 20/15 visual acuity,
time. which is normal for these pilots, and 20/40 degraded visual
The FOV of both systems was rated as adequate. Of the 34 acuity.
Apache aviators, 55% rated the PNVS and TADS resolution Subject pilot ratings indicated that low speed and hover
as inadequate and 75% felt that improving resolution took flight can be performed with reasonable workload using a 23
precedence for a design improvement. Although image quality by 38 degree FOV with normal visual acuity. Also, when acu-
was a problem in Saudi Arabia, 60% of the 66 ANVIS aviators ity was degraded, increasing field of view resulted in little
felt that improving FOV should take precedence based on improvement in pilot ratings.
their total flight experience; another 15% felt that improving The effects of FOV and limiting resolution on flight han-
FOV and resolution should take equal priority. dling were explored in two flight experiments performed by
the Army’s Communications and Electronics Command in the
late 1980s (10,11). Direct-view goggles were built to provide
Flight Experiments various combinations of FOV and resolution. These goggles
The flight experiment results can be summarized as follows. are similar to ANVIS except they do not incorporate an image
With normal eyesight acuity, performance improves with intensifier and are used during the day only.
FOV up to a plateau between 40⬚ and 80⬚ depending on flight Pilots using these goggles were asked to fly preplanned
maneuver. However, degraded visual acuity strongly affects NOE and contour flight profiles. Hover and lateral flight tasks
these results. Once a minimum FOV of about 40⬚ is achieved, were also evaluated. In both tests, trail runs were flown with-
performance is a strong function of image quality. Holding out goggles to establish baseline performance levels. The air-
the sensor FOV to 40⬚ and optimizing image quality is usually craft used was an AH-1 COBRA Attack helicopter with the
the best design tradeoff. subject pilot in the front seat. The aircraft and flight profiles
Increasing FOV by diverging ocular lines of sight (that is, were selected after consultation with test and user pilots.
both eyes see the center third of the total FOV, but the outer Six subject pilots participated, each flying three trials of
third on each side is seen by only one eye) does not improve each task. Measured data included altitude, airspeed, and
performance and may hurt performance. Although the total head motion. After each trial of each task, pilots answered
FOV is increased, the data indicate that fixations and ocular questions on workload, confidence, and aircraft handling
tracking are limited to the central, overlapped region of the qualities. Table 3 shows the combinations of resolution and
FOV. In some important respects, the sensor FOV becomes FOV flown on a test range at Fort Rucker, Alabama in Febru-
the small, overlapped region. ary, 1987.
Based on pilot assessment of flight trials, a detector dwell The term ‘‘ocular overlap’’ in Table 3 is described as
time (exposure time) of 16 ms is unacceptable in a pilotage follows.
system; a dwell time of 4 ms is not noticeable. Also, image With 100% overlap, both eyes see the whole field of view.
processing delays (the time delay between capture of the im- One technique to enlarge the display FOV while maintaining
age by the sensor and display of the image to the pilot) should
be 33 ms or less. Delays of 100 ms lead to serious flight con-
trol problems. Table 3. FOV and Resolution Combinations Flown in
1987 Experiment
FOV and Resolution Trades. In 1975, the U.S. Army Aero- FOV in Limiting Ocular
medical Research Laboratory performed a flight test compar- Degrees Resolution Overlap (%)
ing standard 40⬚ FOV, AN/PVS-5 goggles to modified goggles Unrestricted Normal eyesight Normal
with a 60⬚ FOV (8). On the basis of the flight conditions, the 40 Normal eyesight 100
limiting resolution of the 40⬚ and 60⬚ goggles was 0.6 and 0.4 40 0.9 cy/mrad 100
cy/mrad, respectively. Participating aviators rated the 40⬚, 40 0.6 100
higher resolution goggle as more suitable for terrain flight. 40 ⫻ 60 0.9 50
60 0.6 100
Also, the 40⬚ goggles were associated with smoother, more
60 0.5 100
gradual control stick movements than the lower resolution,
60 ⫻ 75 0.6 75
60⬚ goggles.
 HELICOPTER NIGHT PILOTAGE 677

high resolution is to partially overlap the two oculars of a Table 4. FOV and Resolution Combinations Flown in
binocular display. With partial overlap, both eyes see the cen- 1988 Experiment
tral portion of the FOV, but only one eye sees each edge of FOV in Degrees Limiting Resolution
the FOV. For example, 50% overlap of a 60⬚ goggle means
40 0.9 cy/mrad
that both eyes see the central 30⬚ of the field of view. The 40 0.4
right eye sees the right 15⬚ of the total field of view, and the 40 0.5 at edge/1.1 at center
left eye sees the left 15⬚ of the total field of view. This tech- 60 0.6
nique lets the optical designer reduce weight and volume by 60 0.3
covering a large total FOV with smaller individual oculars. 60 0.2 at edge/0.9 at center
The test device with 40⬚ FOV and with 0.6 cy/mrad resolu-
tion represents current thermal imager capabilities under
very favorable thermal contrast conditions. This combination
also represents the capabilities of ANVIS night vision goggles lution at the center was also evaluated. Table 4 gives the com-
under quarter moon illumination. With the exception of the binations evaluated in the second test which was flown during
device with 40⬚ FOV and normal eyesight resolution, the February and March, 1988. Four subject pilots participated;
other combinations shown in Tab. 3 represent achievable per- each subject flew four trails of each task.
formance in the 1990s time frame under good thermal con- During this test, goggle configuration did not affect alti-
trast or high light level conditions. tude and airspeed performance. Once the task was defined in
The following observations were made based on the Fort the baseline flight, execution did not vary significantly in
terms of the airspeed or altitude which was maintained. The
Rucker test:
highest workload and lowest confidence ratings were given to
the 60⬚, 0.3 cy/mrad goggle simulators. In this test, the pilots
1. When FOV was held constant at 40⬚, decreasing resolu- consistently selected the higher resolution and smaller field
tion resulted in a substantial increase in altitude, a of view devices over the larger field of view but lower resolu-
slight decrease in airspeed, and significantly poorer pi- tion devices.
lot ratings. If resolution at the edge of a 60 degree device was substan-
2. Decreasing FOV to 40⬚ but retaining undegraded visual tially poorer than resolution at the center, two of the pilots
acuity had a very minor impact on altitude and air- consistently rated the 40 degree field of view goggles higher
speed. Pilot ratings for this combination were slightly even when the 60 degree goggles had equivalent or better res-
below the unrestricted baseline but were better than all olution in the central portion of the field of view. The other
other combinations tested. pilots rated these 40⬚ and 60⬚ devices as equal.
After test completion, the pilots were asked to explain this
3. With the 40⬚ FOV, 0.6 cy/mrad device as a baseline, preference. The response was that, with the 60⬚ goggles, they
increasing either FOV or resolution with fully over- would see an object ‘‘and then lose it.’’ This characteristic of
lapped oculars improved performance and significantly the goggles was particularly bothersome during the 360⬚
elevated pilot ratings. When comparing the 40⬚ FOV hover turn out of ground effect but also affected performance
with 0.9 cy/mrad goggles to the 60⬚ FOV with 0.6 cy/ during lateral flight, NOE, and contour flight. It is likely that
mrad device, pilots had some preference for the wider ocular tracking is important in the performance of all these
FOV but exhibited no change in performance. tasks and that poor resolution at the edge of the field of view
4. Increasing FOV by diverging ocular lines of sight (that would therefore lead to adverse pilot reaction. However, ocu-
is using less than 100% overlap of the images presented lar tracking was not measured during the test.
to the two eyes) did not improve performance when the During 1994, a flight test was conducted to test the hy-
40⬚ oculars were used and caused poorer performance pothesis that using an 18⬚ ocular overlap in a 52⬚ total FOV
with the 60⬚ oculars. The 50% partial overlap of the 40⬚ might result in abnormal eye and head movement patterns
oculars resulted in increased head motion and fatigue. (12). A fully overlapped design was also flown for comparison.
Distortion for the 40⬚ oculars was less than 1%. How- The flight test further determined if the difference would im-
ever, distortion in the 60⬚ oculars reached 6%; high dis- pact pilot performance of the prescribed flight tasks. Flight
tortion will undoubtedly cause image convergence prob- tasks included NOE, contour, out of ground effect hover, and
lems between the two eyes and lead to degraded lateral flight.
performance. On the basis of the eye tracking data collected during the
flight, the partial overlap does constrain the eye at the center
of the FOV and significantly reduces the amount of time that
The FOV/resolution combinations tested at Fort Rucker the eye uses the outer portion of the total FOV. Averaged
represented performance projected to be attainable under fa- across all pilots and tasks, the percentage of eye fixations that
vorable thermal contrast or high light level conditions. A sec- occur outside the central 18⬚ when using partial overlap was
ond test was flown at Fort A.P. Hill, Virginia, to explore the reduced by 60% (p ⫽ 0.0170) as compared to the full overlap
resolution versus field of view trade-off when simulating less (full ⫽ 24%, partial ⫽ 9%). There is no difference between
than ideal thermal contrast or light level conditions. tasks (p ⫽ 0.2836).
The FOV/resolution combinations which simulated less Looking at horizontal eye movement, the mean rms ampli-
than ideal conditions were chosen to make the flight tasks tude across the five subjects for the partial overlap was only
difficult but possible. The potential benefit of trading lower 70% of the rms for the full overlap. This 30% reduction was
resolution at the edge of a sensor field of view for higher reso- significant (p ⫽ 0.0136). No statistically significant difference
678 HELICOPTER NIGHT PILOTAGE

in rms amplitude was found between tasks (p ⫽ 0.5022) or head motion. The pilot will see a blurred image for the same
for the interaction between overlap and task (p ⫽ 0.7769). reason that a photograph will be blurred if the exposure time
The average head velocity for partial overlap increases by is too long for the motion being captured.
12.5% and 6% for contour and NOE flights, respectively. Two pilots flew an AH-1 Cobra from the front seat using
The pilots indicated higher workload and lower confidence helmets and helmet-mounted displays from the Apache heli-
when flying the partial overlap as opposed to the full overlap. copter with a small video camera mounted on the helmet. The
Some subjects reported nausea and fatigue after use of the camera FOV was 30⬚ vertical by 40⬚ horizontal and provided
partial overlap; this occurred whether the partial overlap con- unity magnification through the helmet display. The test
figuration was flown first or second. There was no noticeable camera had a limiting resolution of about 0.5 cy/mrad and
visual problem reported on the full overlap configuration. electronic gating to control the dwell time for each video field.
Overall, these results indicate a change in characteristic Selectable exposure times ranged from 1/60 s (one field) to
head and eye motion when the partial overlap is used. There under a millisecond. The pilot’s visor was down and taped so
is a 10% increase in average head velocity and a significant that he flew solely by sensor imagery. The pilots performed
45% increase in the fraction of time that the head is in mo- hover, lateral flight, NOE, and contour tasks. The flight ex-
tion. The data may suggest that the more frequent head dy- periment was performed in January, 1989, at Fort A.P. Hill,
namics may be substituting for the lack of the ocular tracking Virginia.
which is restricted (60% reduction) when the partial overlap Image blur at 1/60 s exposure time was unacceptable. Blur
design is in use. This appears to be consistent with the hy- was present with either aircraft or head motion, and the blur
pothesis that the eyes do not track across the overlap (binocu- interfered with task accomplishment. With an exposure time
lar to monocular) boundary. of 1/120 s, image blur was noticeable with head motion but
The subjective data suggest that the partial overlap effec- no conclusion was reached regarding impact on performance.
tively segregates the overall 52⬚ FOV into an 18⬚ brighter cen- No image blurring was noted at 1/240 s exposure time.
tral and two dimmer outer regions. This perceived decrease Visual acuity is not degraded for ocular tracking rates up
in brightness and acuity apparently derives from the lack of to about 30⬚ per second, and ocular tracking is probably im-
binocular fusion in the outer regions. The subjects indicated portant during pilotage. The exposure time for each snapshot
that luning at the overlap boundary hid scene cues; they sub- taken by a video camera should be short enough that images
jectively rated the partial overlap FOV as being smaller than crossing the sensor FOV at up to 30⬚ per second are not
the fully overlapped FOV. blurred. Note that acceptable exposure time depends on sen-
It appears that the partially overlapped configuration lim- sor resolution; exposure time should shorten as sensor lim-
its ocular tracking, both because of the perceived loss in im- iting resolution improves.
age quality at the overlap boundary and because of the loss
of binocular fusion as the eye tracks over the boundary. The Impact of Image Processing Delays. In advanced helicopter
partially overlapped FOV configuration provides a function- pilotage systems, digital processing will be used to enhance
ally smaller FOV than the fully overlapped configuration. imagery and add symbology. Digital processing adds a delay
An experiment conducted in 1996 evaluated the impact of between when the image is captured by the sensor and when
field of view on precision flight maneuvers (13). Subjects flew it is seen by the observer. This kind of delay is not present in
with FOV restricted to 40⬚ vertical and 20, 40, 60, 80, and currently fielded systems; the impact of this delay on flight
100⬚ horizontal. Normal eyesight acuity was not degraded. performance is unknown. A flight test was conducted to quali-
Maneuvers included pirouette, hovering turn, bob-up and tatively assess the performance impact of delaying pilotage
down, precision landing, acceleration and deceleration, and video (14).
slalom. Performance measures included accurate aircraft po- Two aviators participated in the test and alternated as
sition and heading, head movement, pilot rating of flight han- subject and safety pilot. The subject pilots wore Apache hel-
dling qualities, and pilot rating of visual cues. mets and viewed a helmet-mounted camera through the
Most of the measured data showed a general increase in Apache helmet-mounted display. The camera and display pro-
performance with larger FOV. Flight data indicated that per- vided a 30⬚ vertical by 40⬚ horizontal, unity magnification im-
formance improves with FOV up to a plateau between 40 and age to the subject pilot. During the test, a cloth was draped
80⬚ depending on the flight maneuver. Subjective ratings of over the subject’s visor so that all visual cues came from the
flight handling and visual cues increased with FOV up to a helmet display. A video digitizer provided a variable delay
limit of 60 to 80⬚ depending on task. On the basis of all the between camera and display. All flights were in daylight and
collected data, it was the researcher’s opinion that the great- good weather.
est overall performance gain occurred prior to the 60 to 80⬚ The project pilot established baselines for several, aggres-
FOV range under the conditions tested. sive flight maneuvers using normal day, unaided vision. The
maneuvers included rapid sidestep, pop-up, longitudinal ac-
Image Blur Due to Head and Sensor Motion. A flight test was celeration and deceleration, rapid slalom, nap-of-the-earth,
conducted to determine suitable exposure time for a staring and contour flight. After practicing unaided and with the sen-
camera operating at the standard video frame rate (11). Cam- sor hardware set for zero delay, the subject pilots repeated
eras which use ‘‘staring’’ detector arrays are being considered the maneuvers with the video delay increased after each iter-
for use in night pilotage aides. Most staring sensors use detec- ation of the task set. Test results are based on subject and
tor dwell times equal to the field or frame time of the imager, safety pilot assessments of flight performance.
typically either the 60 Hz video field time or the 30-Hz video On the basis of the qualitative assessment of these two
frame time. In a pilotage sensor, however, considerable image pilots, there appears to be no performance impact from a 33
movement can occur in a video field time due to aircraft and ms image processing delay.
 HETEROGENEOUS DISTRIBUTED COMPUTING 679

Delays of 100 ms or more impaired the subject pilot’s abil- 7. D. Wood, Validation of the Night Vision Requirements for the Army
ity to make stable, aggressive maneuvers. All hover tasks Scout and Attack Helicopter Program, Fort Belvoir: U.S. Army
were more difficult; sometimes a stable hover could not be Night Vision Laboratory, Experiment 43.7 Phase II, 1974.
achieved. Alternate strategies were developed for NOE and 8. M. Sanders, Aviator Performance Measurement during Low Alti-
contour to compensate for the image processing delay. The tude Rotory Wing Flight with the AN/PVS-5 Night Vision Goggles,
Fort Rucker: U.S. Army Aeromedical Research Laboratory, 76-
subjects experienced the feeling that the aircraft motion was
10, 1975.
ahead of the visual scene.
9. R. Hoh, Investigation of Outside Visual Cues Required for Low
On the basis of this limited flight test, processing delays of
Speed and Hover, AIAA Paper 85-1808-CP, 1985.
up to 33 ms cannot be sensed by the pilot and appear to have
10. D. Greene, Night Vision Pilotage System FOV/Resolution Tradeoff
no impact on flight performance. However, with an image pro-
Study Flight Experiment Report, Fort Belvoir: U.S. Army Center
cessing delay of 100 ms, the pilot senses that aircraft move-
for Night Vision and Electro-Optics, NV-1-26, 1988.
ment is ahead of the displayed image. During these flights,
11. R. Vollmerhausen and C. Nash, Design criteria for helicopter
and without prior training with delayed imagery, the 100 ms
night pilotage sensors, Proc. Amer. Helicopter Soc., 45th Annu.
delay led to significant flight control problems. Forum, Boston: 1989.
12. T. Bui, R. Vollmerhausen, and B. Tsou, Overlap binocular field-
of-view flight experiment, SID Digest, XXV, 306–308, 1994.
EVALUATION 13. L. Haworth et al., In-flight simulation of field-of-view restrictions
on rotorcraft pilot’s workload, performance and visual cueing,
Current night pilotage sensors like the ANVIS image-intensi- Proc. Amer. Helicopter Soc., 52nd Annu. Forum, Washington,
fied goggle and the PNVS thermal imager provide a signifi- DC, 1996.
cant capability to fly helicopters at very low altitudes in order 14. L. Biberman (ed.), Electro-Optical Imaging Systems and Modeling,
to hide behind hills, trees, and other terrain objects; this ca- Chapter 26, ONTAR Corp., North Andover, MA, In press.
pability enhances the survivability of tactical helicopters on 15. R. Vollmerhausen, T. Bui, and B. Tsou, The affect of sensor field
the modern battlefield. The availability of heads-up aircraft replication on displayed imagery, SID Digest, XXVI: 667–670,
status symbology, that is, symbology superimposed on the 1995.
night vision imagery, is a critical feature of these pilotage sys- 16. G. Robinson, Dynamics of the eye and head during movement
tems. Further, aviators report that their ability to perform between displays: A qualitative and quantitative guide for de-
night missions is greatly enhanced when both image-intensi- signers, Human Factors, 21: 343–352, 1979.
fied and thermal imagers are available on the helicopter. 17. M. Sanders, R. Simmons, and M. Hofmann, Visual Workload of
Flight experiments and the results of user surveys provide the Copilot/Navigator during Terrain Flight, Fort Rucker: U.S.
guidelines for design improvements. NOE and contour flight Army Aeromedical Research Laboratory, 78-5, 1977.
can be accomplished with reasonable workload using a pilot-
age system with 40⬚ FOV and 0.6 cycles per milliradian lim- RICHARD H. VOLLMERHAUSEN
iting resolution; this resolution provides the pilot 20/60 visual U.S. Army Communications and
Electronics Command
acuity. Improving either FOV or resolution beyond these val-
ues will lessen pilot workload and lead to increased confi-
dence. However, since the ability to resolve scene detail is
important for terrain flight, night sensors should have suffi-
cient sensitivity to provide 0.6 cycles per milliradian resolu-
tion under low thermal contrast or low scene illumination
conditions. In advanced systems, this minimum level of image
quality should not be traded for increased field of view.

BIBLIOGRAPHY

1. Anonymous, Lighting, Aircraft, Interior, Night Vision Imaging

System (NVIS) Compatible, MIL-L-85762A, 1988.
2. D. Newman, ANVIS/PNVS Comparison Flight Test, Fort Belvoir:
U.S. Army Night Vision and Electro-Optics Laboratory, 1982.
3. C. Nash, AH-64 Pilotage in Poor Weather, Fort Belvoir: U.S. Army
Center for Night Vision and Electro-Optics, NV-12, 1987.
4. R. Vollmerhausen, C. Nash, and J. Gillespie, Performance of AH-
64 Pilotage Sensors during Reforger 87, Fort Belvoir: U.S. Army
Center for Night Vision and Electro-Optics, NV-1-30, 1988.
5. T. Bui and J. Gillespie, Night Pilotage Sensor Field Assessment,
Fort Belvoir: U.S. Army Center for Night Vision and Electro-Op-
tics, NV-91-4, 1990.
6. G. Youst, J. Gillespie, and S. Adams, Desert Storm’s Night Vision
and Electro-Optical Equipment Suitability Survey, Fort Belvoir:
U.S. Army Night Vision and Electro-Optics Directorate, AMSEL-
NV-0099, 1992.