@physixguru, do you mean air resistance on the other side of the boat and friction which keeps the boat in place? If so, couldn't this be addressed by instead putting the speaker on a trolly (with greased axles) on greased train tracks? If the sound has any propulsive effect at all then it's just a matter of cranking the volume up high enough so the force overcomes friction.
@gendou2, I think your objection is perhaps more serious. I'm imagining a simplification of this. Imagine a 0.1Hz wave made by a speaker the size of a sail boat sale. It would puff out in one direction, thus pushing air away and moving the vehicle in the opposite direction. Then the sail (encased in a speaker box) would puff in the opposite direction thus pulling the vehicle back in the opposite direction.
The sound would need to be generated such that the push away was fast and the pull back was slow. If I took normal sound produced by a speaker and somehow doubled the length of time the speaker took to pull back in, what would that do to the sound it produced?
After taking a normal sinusoidal tone at some frequency (let's say 60Hz) and then performing that alteration to the wave (the troughs would be made twice as long as the peaks), what would it sound like? What would happen to a normal song if we made that alteration to the sound waves produced by a song?
If I produced a sound of that sort, is there some way to calculate the amount of propulsive force it produces? The above method would presumably move forward 2 steps for every one step back.