HAYES AND GOUGH: SYNTHETIC APERTURE SONAR: A REVIEW OF CURRENT STATUS 209
is especially true for yaw deviations), the correction cannot
produce diffraction-limited images but it still produced image
reconstructions with significantly better along-track resolution
than that relying on the data from the INS alone [74]. There
were several experimental SAS deployed, mostly purpose built
but at least one, a modified commercial side-looking sonar
from the University of California at Santa Barbara [75]–[78].
The biggest and most significant of the purpose-built SAS was
originally made by Alliant (a subsidiary of Honeywell) under a
DARPA contract which was then bought out by Techsystems
then Hughes Aircraft and now owned by Raytheon Electronic
Systems (Waltham, MA). However, since it was a military
system, it produced few publications in the open literature
[79]–[82]. Meanwhile in Europe, under the umbrella of a Euro-
pean Union Marine Science and Technology (MAST) program,
a low-frequency Acoustic Imaging Development/Synthetic
Aperture Imaging and Mapping (ACID/SAMI) sonar project
lead by Zakharia was producing interesting geophysical re-
sults [83]–[88]. At the conclusion of the program, sadly the
sonar was disassembled and the component parts distributed
to the various contributors [89]. A military-based SAS was
commissioned by the Coastal Systems Station group of the
Naval Systems Warfare Center, Panama City, FL (NSWC-PC).
The actual sonar was developed by Westinghouse (now part
of the Northrop Grumman Corporation, Los Angeles, CA).
The NSWC-PC effort was led by Christoff [90] and the SAS
was part of their mine detection program Mobile Underwater
Debris Survey System (MUDSS). They were the first group to
operate a SAS two-band imaging system [91]–[93]. Northrop
Grumman has since built a long-range demonstration system
(known as the ”slow-speed SAS”) with 25-mm resolution at 300
m and mounted it on the NSWC-PC cable-towed body as well
as on the Applied Research Laboratory at Pennsylvania State
University (ARL/PSU, State College) Seahorse autonomous
underwater vehicle (AUV). ARL/PSU also built SAS for the
NSWC 12.75- and 21-in AUVs [94], [95].
With the cable-towed deployment, the first choice was the
much heavier than water platform with a bridle attachment from
the top of the towfish much as is done with standard side-looking
sonars. This unfortunately resulted in poor yaw stability since
the angular restoring moment to yaw disturbances was low. This
did not matter greatly for side-looking sonars or for single-hy-
drophone SAS systems. However, it was a significant impair-
ment when multiple-hydrophone arrays evolved as any uncor-
rected yaw errors injected high-frequency errors into the raw
data even when the uncorrected yaw errors themselves were
no worse than low-order polynomials. In contrast, the neutral
buoyancy nose-towed fish had excellent yaw stability but cor-
responding poor roll stability which was fine until the devel-
opment of InSAS where uncorrected roll errors produce height
uncertainty.
Most early AUVs and underwater unmanned vehicles
(UUVs) used a fixed position propeller with hydroplanes front
and rear for closed-loop attitude control. Of course, the pro-
truding hydroplanes made deployment and recovery difficult in
a seaway so some AUVs have migrated to the vectored thrust
system of attitude control where the propellor is contained
within a gimbaled shroud. It appears that although the average
attitude of the vectored thrust AUVs could be maintained to
small errors, the instantaneous attitude relative to the set attitude
was more variable than those experienced in the cable-towed
platforms (the control of a vectored thrust AUV is somewhat
akin to that of balancing a broom on the end of a finger). Of
course, for SAS, the instantaneous attitude is as important as the
average attitude but the teething problems with vectored thrust
propulsion on AUVs seem to be less of an issue and high-quality
images from the NSWC-PC are appearing in the open liter-
ature [71], [92], [95]–[97] as well as from others [98], [99].
Meanwhile, the Acoustics Research Group at the University of
Canterbury (Christchurch, New Zealand) had evolved KiwiSAS
II into a neutral-buoyancy nose-towed SAS [100]–[102]. Being
university rather than military based, they had designed and
built a new version of their lightweight, low-cost SAS. Despite
this approach, their home-built transmitting projector had a
bandwidth and efficiency not matched by any commercially
available transducer at that time [103], [104]. They also worked
on fast Fourier domain image reconstruction algorithms and the
efficacy of the motion compensation algorithms [105]–[111].
Lately, they have been attempting to validate the algorithms
using an active beacon system [112]–[114].
Also joining the SAS community was a group led by Grif-
fiths at University College London (UCL, London, U.K.) with
DERA (now part of QinetiQ, Farnborough, U.K.). The UCL
group was mainly interested in porting SAR-based interferom-
etry techniques to SAS [99], [115]–[118] and combining when
appropriate with the experimental group at DERA [119] for
some real-world experiments. A Dutch research group based
at TNO (Delft, The Netherlands) and the Delft University of
Technology (Delft, The Netherlands) has used the French rail-
based SAS [60] to verify the efficiencies of some image re-
construction algorithms as well as to test some path estimation
and MOCOMP techniques [62], [120]–[124]. They have also
trialled ship-mounted SAS operation in conjunction with the
French research group Groupe d’Études Sous-Marines de l’At-
lantique (GESMA, Brest, France) [122], [125]. In Sweden, a
research group at Chalmers University (Göteborg, Sweden) has
made several experiments measuring the along-track resolution
and the efficacy of path estimation and MOCOMP algorithms
[126]–[130].
While many of the early groups driving research into SAS
were either universities or government laboratories, a few
commercial companies developed their own SAS systems as it
became a proven technology. A group within Applied Signal
Technology (Los Angeles, CA), previously a separate company
called Dynamics Technology, Inc., as well as Bloomsbury
DSP (London, U.K.) and Wavefront Systems, Ltd. (Sher-
borne, U.K.) all sell complete software packages with SAS
processing. The most mature of these three groups, AST/DTI
had an earlier background in SAR processing and although
their processing techniques are often clouded by commercial
reticence, they have published some excellent results with their
image of the PBY40 plane in Lake Washington being one of
the best of that particular target [131]–[134]. More recently,
AST/DTI have demonstrated bathymetry using a 175-kHz
InSAS on MacArtney’s (Houston, TX) dynamically controlled
tow vehicle [for the U.S. Office of Naval Research (ONR)], a
Authorized licensed use limited to: TIANJIN UNIVERSITY. Downloaded on May 03,2021 at 06:45:16 UTC from IEEE Xplore. Restrictions apply.