Soon after beginning to work on brain-machine interfaces, Dr. Ojakangas felt compelled to build a robotic arm that is actuated by simulated muscles, in order to gain a better understanding of how neural impulses from the motor cortex in a human could activate numerous muscles simultaneously to achieve the goal of a prescribed arm motion. The robotic arm, named “Son of Toby” after an actual human skeleton housed in the Drury University science division for over 100 years, eventually had six simulated muscles and two degrees of freedom (shoulder and elbow joints). Two Drury University physics department teams flew versions of this arm on the Weightless Wonder aircraft, testing their ability to induce the arm’s hand to trace out various planar trajectories in the absence of gravity. In developing this invention, the assistance of mechanical engineer Joshua Petitt was invaluable. In the equations of motion, the acceleration due to gravity, g, could be set to any value. The team was therefore able to set the value to 9.81 m/sec2 or to zero, and perform subsequent tests of its motion in weightlessness or on earth. The experiments worked, and the arm successfully traced prescribed paths (straight lines and circles) both in the laboratory at Drury University and on the aircraft. The arm even wrote its name in the lab (Son of Toby) on a whiteboard. The history of this project is laid out below:
The original idea was to actuate the arm of a human skeleton model with twelve simulated muscles constructed from servos (representing muscle contraction), elastic components, and pulleys (representing joints). On the left is the model skeleton, and on the right is the same skeleton model with a set of elastic tubes, in appropriate positions, that were to be replaced by the full simulated muscles. The apparatus was to be controlled by a Brainstem microcontroller by Acroname corporation.
The original plan had a ball joint for the shoulder (which has three degrees of freedom) and a simple elbow joint with one degree of freedom. In other words, there were four angles needed to describe the orientation of the arm. However, the equations describing this system turned out to be very complex, and riddled with “singularities” — values of the angles for which the equations “blow up” . The graphics on the left and the right display some of these singularities in the four-dimensional space of the arm’s angles. The answer is to use quaternions, but this arm was just too ambitious. We needed to start with something simpler! Nonetheless, we worked out the theory, which is described in the appendix of this proposal to the Ncircle6g_gwo
ASA Reduced Gravity Student Flight Opportunities program.
The second version of the arm, much improved, was flown in the summer of 2011. Below (left) Preston Julian enjoys weightlessness while testing the arm on the aircraft. Middle: the arm is tested under normal gravity, in the lab. Right: virtual Son of Toby writes his name backwards.
Any questions?