TECHNICAL PROPOSAL
                         January 20, 1988
General Considerations:  The design selected must be strongly 
built, readily available on the used boat market, and well 
ballasted in proportion to its sail area.  That is, the design 
should not rely heavily on moveable crew weight for righting 
moment.  Hull, deck, rig, and appendages must be capable of 
withstanding a knock-down or roll-over.  On most available boats 
this will probably require some reinforcing of the rudder, and 
replacement of spars and standing rigging.  Full keel/attached 
rudder configuration has the advantage of being less susceptible 
to fouling by sargasso weed or other floating objects, and less 
likely to sustain rudder damage.  Fin keel/spade rudder 
configuration (possibly including a bulb keel) has a considerable 
performance and control advantage, especially downwind.  
Size: Several factors determine the minimum feasible size of the 
                1) Must be large enough to carry off-the-shelf 
                   navigation, computing, steering, and 
                   communication equipment without serious 
                   degradation of performance.  The costs and 
                   long lead times associated with 
                   miniaturization are to be avoided. 
                2) Must have sufficient deck area for the 
                   required solar panels for battery charging. 
                3) Rig must be able to support a masthead wind 
                   instrument cluster. 
                4) Backstay must be a reasonable length for a 
                   short-wave radio antennae. 
                5) It may be desirable for the vessel to carry 
                   passengers during testing and de-bugging. 
                6) The model should be visible to other vessels, 
                   so as to avoid risk of collision. 
              Several factors also determine the maximum feasible size:
                1) Vessel must be small enough to rig and test 
                   with minimum lead time. 
                2) Sails must be small enough to withstand 
                   serious abuse in adverse weather conditions. 
                3) Rig must be small to withstand roll-over with 
                   minimum probability of damage. 
                4) Reefing systems should be kept simple, and the 
                   number of sailplan configurations required 
                   during the voyage should be kept to a minimum.   
                5) The probability of being spotted by other 
                   vessels should be minimized, so as to reduce 
                   the risk of the vessel being "rescued" or 
                   otherwise interfered with. 
                6) Vessel should be easy to transport and store. 
                7) Vessel must be acquired quickly and at 
                   moderately low cost. 
                8) The danger, real or perceived, of causing 
                   damage to another vessel must be minimized. 
Preliminary evaluation of the above size parameters suggests a 
vessel size of between 16 and 24 feet.  Candidate stock designs 
       Full keel designs:  Cape Cod Bull's Eye (16 ft, 1350 lb.)
                           Cape Dory Typhoon (18 ft, 2000 lb.)
                           Pearson Ensign/Electra (22 ft, 3000 lb.)
       Fin keel designs:   Zypher (16 ft, 500 lb.)
                           International 110 (24 ft, 1000 lb.)
                           Wylie Wabbit (24 ft, 700 lb.)
Steering will be controlled by an electrically-driven compass- 
referenced autopilot.  Wind vane type self-steering is ruled out 
because of size and weight, vulnerability to wave damage, 
difficulty in making automatic adjustments, susceptibility to 
fouling the pendulum with floating objects, and the difficulty in 
providing redundancy in the event of failure. 
Several suitable off-the-shelf autopilot systems are available, 
specifically the Autohelm 2000 or 3000.  Although these units are 
reasonably reliable, some redundancy is necessary.  One possible 
actuator configuration is to install two or three tillers on the 
rudder stock (all below deck level).  Each tiller will have its 
own steering actuator.  An electrically controlled clutch device 
will couple only the active tiller to the rudder, and only the 
actuator on that tiller will be powered.  Similarly, two or three 
control units will be available.  On-board diagnostic software 
will determine  which tiller/actuator set and which control unit 
should be switched on. 
The model will carry one satellite navigation unit, plus one 
additional satellite unit or one LORAN unit.  Both of these 
systems are available off-the-shelf for the recreational boating 
market, complete with a standardized data output interface.  
There will also be two remote-reading compasses, and two 
knotmeter/log units.  These devices will send data to the on-
board computer for maintaining a dead reckoning position, and/or 
used in conjunction with the SATNAV or LORAN units to utilize 
their built-in DR capabilities. 
The on-board computer software will make course decisions based 
on wind direction and speed measurements, the current vessel 
position, and the pre-programmed or transmitted routing strategy. 
In addition to the navigation instruments (knotmeter, log, and 
compass), the model must also be capable of measuring wind speed 
and direction.  This could be a standard yacht masthead 
instrument cluster, or possibly a heavier commercial anemometer 
and remote direction indicator.  Although the yacht type unit is 
less expensive, lighter, and consumes much less power, these 
devices have a poor reliability record and cannot be expected to 
survive a roll-over.  
To allow for the possible loss of wind measurement capability, 
the model will also measure its heel angle.  This information, 
combined with course, speed, sail trim settings and reef 
configuration, will allow the on-board software to estimate wind 
speed and direction with reasonable accuracy.  A short wind test 
maneuver, performed hourly, might be used to improve the accuracy 
of this wind estimate if the masthead cluster is lost or 
The primary communication system will be via short-wave radio, 
pre-tuned to transmit and receive "packets" of digital 
information at selected times.  Even at a relatively slow data 
transmission rate and with sveral repetitions, these 
transmissions will only take a few seconds for each contact  
between the model and the land-based communications station.  A 
range of frequencies can be used to maximize the probability of 
making a successful contact under varying propagation conditions.  
It will be desirable to use one ham station on the east coast of 
North America, and one station in England,  linked by telephone 
to the central control and information center.  
Satellite bulletin boards may also be extremely useful, but 
antennae aiming will not be possible from the vessel. 
Contacts will be attempted on the order of twice daily.  
Transmissions from the model will include: 
     1) Present position from DR.
     2) Last SATNAV/LORAN fix.
     3) Present vessel speed and heading.
     4) Present wind speed and direction.
     5) Heel angle, sail trim settings, and reef configuration.
     6) Battery condition and charge/discharge current.
     7) Time history of items 3,4,5,6 since last contact.
     8) Results of diagnostic routines.
Transmissions to the model will include:
     1) Routing strategy modification, based on weather information 
        available on shore.
     2) Modification of time and/or frequency of next contact 
        attempt, if necessary for propagation conditions or security.
     3) Initiate switch to arrival mode, initiate real-time control 
        with specified parameters.
     4) Modify arrival waypoint if necessary.
In addition to long-range communications, a VHF radio 
(approximately 25 mile range) may be installed as an 
interference-avoidance aid.  This radio would be activated once 
the model is at sea, and continuously monitor channels 13 and 16.  
If a call from another vessel is detected, the radio would return 
a taped message explaining the nature of the voyage, along with a 
request to report the model's position.
Sail trim:  In order to minimize the possibility of tangles and 
fouled rigging, running rigging must be kept as simple as 
possible with 1:1 sheets and control lines where possible.  
Linear actuators, similar to those used for the autopilot tiller 
control, will be used to adjust sheets.  Note that the duty cycle 
of an actuator used in this way is extremely intermittent, and 
their reliability should be far greater than that of the 
actuators used for steering.  The mainsail will be controlled by 
a single-part sheet and one or two actuators.  The jib will be 
most likely be a conventional non-self tacking working jib, with 
an actuator controlling each sheet independently.  This 
arangement will probably offer the least possibility of a 
tangle, maintain good trim through the widest range of wind 
speeds and angles, and work well with roller furling.  A single-
sheet arrangement with a boom or tacking track might still be an 
attractive alternative, however. 
Reefing:  The jib will be roller-furled around a rigid luff-
support spar.  Although luff-support reefing/furling systems are 
subject to reliability problems on larger vessels, the approach 
here will be to use a grossly oversize system.  There will be 
only two positions for the jib: furled and unfurled.  The furling 
spool will be driven by a continuous chain drive from a small 
electric motor. 
There will be one fairly deep area reef in the mainsail.  A 
linear actuator will slack the halyard, tighten the reef clew, 
and tighten the reef tack simultaneously.  No lacing lines will 
be used. 
In order to insure smooth working of the mainsail luff, steel 
lugs will be sued at the inboard end of each batten.  A 
lubrication system will introduce lubricant into the track.  In 
moderate weather, the reef will be exercised about once a day to 
prevent salt build-up or freezing of moving parts. 
All actuators and servo-motors for trim and reefing will be 
located either inside the main boom, or in a special shallow 
compartment located just below the main deck, sealed from the 
rest of the model's interior.  Because a small amount of water 
will enter this compartment through the ports where control lines 
penetrate the deck, here must be provision for draining or 
pumping this water overboard. 
Electrical power will be provided by several large lead-acid deep 
discharge type storage batteries, maintained by solar panels on 
deck and possibly also on the model's starboard side.  
Approximately 100-150 watts of peak rated charging power will be 
required for a summer crossing, which entails approximately 20-30 
square feet of solar panel area (using off-the-shelf marine solar 
charging panels).  For a winter crossing, it is estimated that 
approximately three to four times the peak rated charging 
capability will be needed, unless other  power-economizing 
measured are taken.  A low power mode will be available to the 
model, with reduced steering input, less frequent sail trim, and 
possibly delayed or abbreviated radio contact. 
The on-board computer will probably be an off-the-shelf MS-DOS 
laptop computer, selected for reliability in adverse or more 
computers will be installed for redundancy, although only one 
will be environmental conditions.  Two operating at any one time 
under normal conditions. 
Departure and arrival will require real-time steering control 
from an escort vessel.  This will be accomplished by sending 
course instructions to the autopilot by means of conventional 
designated radio control frequencies.  If the model reaches the 
designated arrival waypoint before the escort vessel is on 
station, then the model will heave to and send an intermittent 
homing signal on VHF frequencies.