A step-by-step approach for perfecting intelligent myoelectric controller algorithms

The evolution of the intelligent algorithm that is capable of preferentially reallocating the action selection times in a myoelectric controller for multiple degree of freedom prosthesis, is the theme of this paper. Here thirteen algorithms, their advantages and disadvantages are discussed and the reasons for modifying them leads us to improved versions of the algorithms, which takes us closer to the ideal algorithm for intelligent myoelectric controllers. The need for self preferential time allocation arises because of two prime reasons. Firstly, the programmer of the myoelectric controller is not going to be around while the user uses the prosthesis and hence will not be able to reprogram the device. Secondly, application specific myoelectric controllers cannot be manufactured for every customer in reality. The embedded intelligence allocates the least time to those actions that are performed the most number of times. This is done with a combination of spatial and temporal coding, so that temporal priorities can be reassigned preferentially to spatially coded signals. By using intelligence embedded algorithms in the place of the existing ones which have to be allocated the action selection pulse times beforehand, we eliminate the need for individually programmed myoelectric controllers.