It is possible to automate a simple yacht. There is a project underway to build small approximately 1 1/2 meter sailing craft capable of crossing the Atlantic without anyone on board. I saw a prototype of one of these and it was really amazing.
There was also a project done in Germany where a 10 meter racing sailing yacht was purposely constructed so that it could be fully instrumented so that all movement and all corresponding forces on the yacht could be measured and documented.
But both of those experiments are very different than creating a training vessel which was fully instrumented.
Having taught at least 100 people how to sail over the years, ideally, a good training vessel is very responsive; so that the student can make an adjustment and feel the change. It allows the beginner to sense the cause and effect of each action so it becomes ingrained in the student's mind what any particular change in sail trim or course is likely to do to the behavior of the boat.
Although somewhat contradictory, the ideal training boat should not scare the daylights out of a beginner either.
Ideal training boats for adults have evolved over time, but these days they typically look something like a Colgate 26, J-80, Sonar 23, or J-22, in other words a 23 foot (7 meter) to 26 foot (8 meter), fin keel, spade rudder, fractionally rigged, moderately light displacement sloop with a tiller and a large cockpit.
Boats this size allow a small number of students per teacher, and having the treacher right there is mostly what it takes to alleviate, or at least minimize the beginner's fears. This is a great venue to allow someone to experience sailing easily and to teach someone sailing basics quickly.
Then there is the question about using Trimarrans for traing purposes. I really like trimarrans for certain purposes. I have owned one and have designed one, and really got a kick out of them. But Trimarrans make a terrible platform to teach sailing, and make a really awful candidate for automation. For their own safety at sea they need to be light enough to disburse the force of the wind by accellerating rather than heeling (or have a very small sail area or a combination of both).
Trimarrnans come in several flavors but taken individually none make sense as a automated training boat.
There are extremely high performance trimarrans. They are very fast, but take a huge amount of skill to sail well. Sailed properly they can win an America's Cup or break a record circumnavigating the world, but these are bleeding edge designs which take extremely high levels of skill to even keep upright and together.
There are mid-range racer-cruisers like the Corsairs, Farriers, Telstars and Dragonflys. These give up a little performance for a higher degree of safety, but they would still be a pretty poor platform for teaching sailing since there is a less visceral connection between cause and effect, and they still require a pretty high level of skill to sail safely in changing conditions. They would also be pretty hard to automate as well. Their inherent speed means that there are comparatively smaller changes in measurable apparent wind angles making accurate approximations of instanteous true wind directions harder to detect, and issues of gradient effect when accellerating out of a tack or gybe more significant. If the computer gets it wrong, there is a higher chance of breaking the boat or hurting a student. But also, to some extent, the faster motion of a trimarran and the need to travel more distance from tramp to tramp more quickly, makes it harder for a student to know where and how to move around the boat, and without an instructor saying "now" and pointing it can be confusing. (I almost lost my first wife overboard on several tacks when she was trying to get off the tramps.)
Then there older style cruising tris. They are more docile and forgiving, but they generally did not tack all that reliably and so there would need to be a computer regime that could detect being in irons, reversing the sails and helm as the boat began moving backwards. But even if these an older style tri could be automated, but they would be the worst of the bunch as a training venue and so not really suited to your objectives.
You say this is an academic exercise in industrial design and service: I would suggest that in any design process, it is helpful to understand what has been done before and why, And only then try to define what you want to improve about the current solutions, which should lead you torwards your solution.
In this case, you appear to have started out chosing a solution; one that is an improbable solution to the problem, which may have emerged out of a misunderstanding of the design problem and the limitations of the currently available technology, and are trying hard to find a problem to fit your solution to.
In reality Trimarrans and automation are wildly expensive technology (that cannot really solve the scariest part of being a beginner on a boat, colisions and docking manuevers) that make little sense if your goal is ease a beginner into sailing. You might instead look at some of the graduated solutions such as the Laser radial rig. Or how junior sailing programs take total novices and get them aclaimated to sailing and then work them up as their skills develop.
You might also look at the O'Pen Bic
YouTube - O'pen BIC, Energized Sailing which is just one approach, albeit a pretty aggressive one.....
Respectfully,
Jeff