Study of Myoelectric Prostheses Based on Fuzzy Control and Touch Feedback

There are two important problems in the system of myoelectric prostheses with tactile sense. One is how to obtain the characters of myoelectric signal (MES) and based on it, implement motion pattern recognition. The other is the method of the bionic control with touch feedback. Many experiments indicate that traditional method of MES processing can not avoid the influence caused by the individual difference. In order to reduce this disadvantage, the unmeasured noise mixed in the gathered MES is eliminated through signal pre-processing firstly. The crosspower spectrum coefficient is extracted from MES, using as its eigenvalue. Then, basing on touch feedback, design a fuzzy arithmetic to control the mobile interval of the myoelectric prosthetic hand. The experimental result indicates that, by using this method, it can not only get a higher recognition rate of motion pattern, but also realize the stable grasp and diminish the possible damage to grasped objects.