1. The vehicle operated without any mechanical or electrical failures.
2. The vehicle was controlled with real-time motor controls and per-
formed the maneuvers necessary to submerge and follow the lip of the elliptical
disk at TRANSDEC.
3. A good quality television image, received over the umbilical cable,
was recorded on video tape.
4. Super 8-mm motion pictures were taken by the vehicle; they were acti-
vated remotely by the surface operator.
Similar results were obtained in the bay tests, except that visibility was
poor and television and motion picture coverage of the ocean floor was limited
accordingly. The controllability of the vehicle is amazing in that the ve-
hicle is long and narrow and side drag prevents the use of differential thrust
(reversing one motor) during turns. However, even with this limitation, a
minimum turn radius of approximately 9 ft was observed.
Only minor problems were encountered, but they indicated the direction
for improving the vehicle design which ultimately dictated the vehicle con-
figuration described in the section on system description.
VEHICLE TRIM. Although two new 3-in battery bottles spanning the length of
the vehicle had been designed, they were not yet fabricated and installed on
the submersible at the time of these tests. Thus, the batteries again caused
a bad list, to one side of the vehicle. This caused an undue amount of delay
in trimming the vehicle in the water. It took 16 lb attached to the forward
starboard bottle and an additional 7 lb in the midsection of the vehicle on
the starboard side to trim the vehicle. For the saltwater operation, the
vehicle required an additional 12 lb attached to the center runner at the
midsection of the vehicle. The vehicle trimmed out at about 1 lb positive
buoyancy. Now that the new 3-in battery bottles have been fabricated and
installed, the time required to trim the vehicle should be greatly reduced.
JOYSTICK CONTROLS. Direct control of the vehicle during this series of tests
was achieved through operator interaction with the cursor controls and the "D"
and "U" keys on the console keyboard. The use of these controls was a "quick
and dirty" approach to operator interaction. A fair amount of concentration
was required to remember which keys performed which function. This approach,
therefore, has been replaced with two joystick controls mounted on each side
of the console keyboard. The right-hand joystick controls continuous forward/
reverse and right/left movements of the vehicle. The left-hand side dimen-
sional stick provides proportional depth control to the vertical thruster.
This approach has proven much more satisfactory in the laboratory and is now
the new permanent means of analogic control of the vehicle when used in pro-
jection or real-time control modes.
CHARGING TIME. Much operational time was lost during these tests because of
the time required to charge the batteries. A more modular approach has been
adopted because of this experience. A set of extra battery bottles has been
fabricated, allowing the operation to continue by changing the bottles in the
field and recharging the first set while using the second set.
CONCLUSIONS AND RECOMMENDATIONS
The tests validated the achievement of the design goals discussed on
pages 9 and 10 of this report. In addition, the tests indicated that the
supervisory-controlled vehicle configuration allows a reliable, flexible soft-
ware structure which provides a testbed for both projection and replacement of
manned concepts. The modular software and hardware architecture concepts were
not only realized, but they have helped finalize the designs which overcame
problems encountered in the vehicle's development history.
Of the problems encountered during the testing phase, man/machine inter-
action stood out as an area which could always be improved. It was difficult
in the field to visualize, communicate, and verify quickly the a-priori
derived, data base information required to command the vehicle to execute a
desired programmed trajectory. An editor- type program was required and aug-
mented with a means of plotting the planned trajectory. Analogic displays
were developed when possible to help the operator visualize the status of the
operation at any given moment to aid real-time vehicle control in the field.
This entire man/machine interface can be further improved in the future.
Speed of interaction and ease of understanding the computer output information
are the key areas for improvement. Similar man/machine improvements are fore-
seen as desirable as part of the manipulator package: A simulated master arm
would allow analogic communication with the vehicle manipulator and special
sensors and a moving cursor display would provide a means of keeping track of
the manipulator claw through a low bandwidth communication channel. Such
approaches are highly recommended in the design of future remotely or
supervisory-controlled undersea vehicle systems.
Although the EAVE WEST testbed is far from being a truly autonomous
undersea system, the approach taken to combine projection modes of operation
with replacement (or autonomous) modes of operation has proven to have several
advantages. First, advanced autonomous concepts are being approached one step
at a time. Second, the operator can always "take over" when an autonomous
operation experiences problems or is completed. Third, through experience, it
is easy to identify primitive autonomous operations which would be useful even
in undersea vehicles which are primarily operated as remotely controlled ve-
hicle systems (RCVS) in projection mode. The eventual autonomous configura-
tion might be configured from a hierarchy of such primitive operations.
It should be remembered that shape of the frame for the EAVE WEST vehicle
can easily be changed to fit special requirements and missions. This is par-
ticularly important in commercial applications where it becomes quite cost
effective to transfer the motors and electronic bottles to a new frame to
achieve large differences in the performance of the vehicle in speed, maneu-
verability, or static stability. For example, it is possible to build a
square or rectangularly shaped frame to make a much more stable platform for
use with the manipulator configuration. On the other hand, speed could easily
be traded for mechanical flexibility by packaging the entire pipeline inspec-
tion configuration into a cylindrical fairing using the fairing itself as the
mechanical frame and gussets to hold the various bottles, sensors, and effec-
tors. In either case, the present modular frame would simply be set aside.
The next step in the design of unmanned free- swimming submersibles should
incorporate some more advanced form of artificial intelligence. As visual
sensors are added to the EAVE WEST submersible, consideration should be given
to the derivation of control information from these sensors by the vehicle
computer. Scene analysis techniques or perhaps image processing of the video
signal should help make this possible. The vehicle has proven itself to be a
fairly reliable versatile testbed. It is easily adaptable to the test and
evaluation of a wide variety of artificial intelligence concepts and it is
recommended that work continue in this area.
1. R. Frank Busby Associates, "Underwater Inspection/Testing/Monitoring of
Offshore Structures," February 1979 (Commerce Contract 7-35336).
2. S. Y. Harmon, "Application of Robot Technology to the Marine Tactical
Environment." June 1980 (in publication).
3. Naval Undersea Center, "Why Man?" Technical Note 953, by H. R.
Talkington, February 1973.
4. Naval Undersea Center, "Manned and Remotely Operated Submersible Systems:
a Comparison," Technical Publication 511, by H. R. Talkington.
5. R. A Geyer, "Submersibles and Their Use in Oceanography and Ocean
Engineering," pp. 61-95, Elsevier Scientific Publishing Co: New York,
6. "Experimental Autonomous Vehicle Program EAVE; Development of Unmanned,
Untethered Submersible Technology for Inspection Tasks," prepared for
USGS Research and Development Program for Outer Continental Shelf Oil and
вЦ† Gas Operations.
7. D. R. Blidberg, E. E. Allmendinger, and N. Sideris, "The Development of
an Unmanned, Self-Controlled, Free-Swimming Vehicle," Proceedings of the
Offshore Technology Conference , 1978.
8. W. R. Ferrell and T. B. Sheridan, "Supervisory Control of Remote
Manipulation," IEEE Spectrum , Volume 4, No. 10 pp. 81-88, October 1967.
9. S. J. Cowen, "Fiber-Optic Transmission System Pulse Frequency Modula-
tion," Proceeding of Oceans 79 , pp. 253-259, September 1979.
10. Naval Ocean Systems Center, "Fiber-Optic Communication Links for Unmanned
Inspection Submersibles," by S. J. Cowen, M. Kono, and P. J. Heckman
(technical report in publication).
11. Naval Ocean Systems Center, "FY 77 Subsea Slow Scan Acoustic Television
(SUBSAT) Tests," Technical Report 217, by A. Gordon, March 1978.
12. T. L. Brooks and T. B. Sheridan, "Experimental Evaluation of the Concept
of Supervisory Manipulation," 16th Annual Conference on Manual Control,
MIT Cambridge, Massachusetts, 5-7 May 1980.
13. P. Heckman, D. Yoerger, T. B. Sheridan, "The NOSC/MIT Submersible Manipu-
lator: an Experiment in Remote Supervisory Control of a Microprocessor
Based Robot," Session F10P on Robotics, International Conference on
Cybernetics and Society, Cambridge, Massachusetts, 10 October 1980.
14. P. J. Heckman and H. B. McCracken, "An Untethered, Unmanned Submersible,"
Proceeding of Oceans 79 , pp. 733-737, September 1979.
APPENDIX A: FREE-SWIMMING SUBMERSIBLE ASSEMBLY