Prosthetics are often controlled by the surface electromyogram (sEMG) of muscles remaining in the limb to which the prosthesis is attached. Typically this is achieved using the on-off 1 degree of freedom approach. A major challenge for next-generation prosthetics is to design a proportional control scheme that allows for control of a number of degrees of freedom.
We hypothesise that the flexibility of the nervous system is such that individuals may supersede innate biomechanical constraints on upper-limb muscles and joints, and that this ability may be exploited to develop abstract muscle groupings for the purpose of myoelectric-control.
To demonstrate the feasibility of this approach we use co-contraction of muscles in the forearm which typically counteract one another. Calibrated muscle activity, as measured by sEMG, determines the position of a 2-D cursor in the MCI control space. The position of the cursor in the MCI control space is then used to select what shape a hand, either virtual or artificial, should be moved into.
Presenter: Dr Matthew Dyson