A neural network controller for a prosthetic limb

Recent advances in computing technology along with improvements in signal processing techniques provide exciting opportunities for the development of highly functional rehabilitative aids for the disabled This paper reports the use of an artificial neural network (Jordan's Sequential network) for use in a myoelectric prosthetic arm to facilitate fault-tolerant pattern recognition and simple sequential tasks. Initial experiments with this technique have yielded promising results and indicate that the technique has a very good potential for an effective rehabilitation of the functionally disabled in terms of their employability, independence, and self-esteem.