12V isn't much drive for the transducers, and the circuit you forgot to post might not be coupling it efficiently to the transducers.
The paper you linked mentions a 35V power source for the transducer driver, which if efficiently coupled, is in theory about 9 times the energy of a 12V source. Furthermore, their phased array focused the energy at a point.
I'm not surprised you didn't feel anything, so consider reproducing the work in the paper you linked.