ABSTRACT
People are often in situations in which they must grasp and manipulate objects with hands that have no sensation. In some cases the insensate hands are their own, such as when diabetes mellitus or Hansen's disease has severely damaged the peripheral nerves; or the insensate hands may be mechanical, such as a motorized prosthetic gripper. It was, in fact, the development of the myoelectrically controlled prosthetic hand shortly after World War 11 (Reiter, 1948, cited in Körner, 1978) that brought the problem of manipulation without sensation into focus (Wiener, 1951; see also Sueda & Tamura, 1969). Prior to that time, upper-extremity prostheses were controlled by a cable connected to a shoulder harness at one end and connected to the prosthetic hand (typically a Dorrance hook) at the other. The user pulled on the cable by shrugging the shoulders, opening or c10sing the hand. Some sense of grip force was provided by the tension developed in the cable. In the myoelectric hand, in contrast, a user generates an electrical signal (EMG) by contracting one or more musc1es. The EMG is
hand. In this case, there information hand
Artificial sensory feedback, that is, presentation of missing sensory information through an alternate sensory channel, can improve the dexterous control of an otherwise insensate prosthetic hand. Meek, Jacobsen, and Goulding (1989) found that grasp-force feedback via a mechanical indenter on the upper arm improved the success rate for normal subjects using a modified myoelectric prosthesis to lift a heavy object without breaking it. Similarly, Kawamura and Sueda (1968) reported that electrocutaneous feedback of grasp force and finger span in a myoelectric hand increased the success rate for picking up quail's eggs from 60070 to 100%. Patterson and Katz (1992) found that feedback via apressure cuff allowed subjects to match grasp forces more accurately, and Mann and Reimers (1970) showed that a vibrotactile code for elbow angle improved the accuracy of elbow flexion with the Boston arm.
