Well, a (cross) bow is mechanical superior for mainly two reasons:

1) The only part that is moving with the speed of the projectile is the center of the string - which is very light.

2) The geometry of the flattening triangle between the nock of the arrow and the bow tips makes the mechanical advantage between the bow and the arrow change along the acceleration path of the arrow. This allows re-gaining much of the mechanical energy caught in the movement of the bow arms and putting it into the arrow. It basically slows down the bow arms and (further) accelerates the arrow in return.

Torsion springs are superior to leaf springs (like a bow) as they make better use of the springs circumference. But the benefits mentioned above make up for more than the difference.

Having a torque driven motor is possible. The torque of a motor is proportional to the current. So you'd need both a current regulation and a rpm sensor to detect when to motor stalls out.

I would rather not go down this route but instead measure the distance traveled. This can be done much more easily by counting rotations - either by back EMF or by using a photoelectric sensor.

If you are looking for a mechanically efficient design you should consider cross bows.

You also should think about how to reload the device automatically. Having varying draw lengths makes this a lot more complicated.

Do you already have a scale / range in mind?

Hey! Sorry for the later reply.

1) I was thinking about that actually, spring vs elastic band (crossbow) system.
I thought in the end that a spring driven system might pack a bigger punch and takes up less space, such that I can even add more springs and upgrade to springs with higher spring rates.

2) The reload mechanism.... I'm definitely planning on automating it, as the whole aim of the project is to make the whole thing autonomous. But for an initial prototype, I want to focus on the mechanics itself, taking it one step at a time :)
Any suggestions? I was actually thinking of having a torque driven motor where I can just set the torque electronically to a specified rate. That would be the automatic winding. as for ammunition... I haven't really though about it, but it's definitely something I plan on experimenting.

3) I don't have a scale in mind at the moment. I'm kind of just starting small and building up the more intimate I get with the device.