Design, Build, and Test of a Permanently Implanted Prosthetic Hand

The feasibility of a permanently implanted prosthetic hand was evaluated from both an internal biocompatibility and exterior mechanics point of view. A literature review of the issues involved in permanent implantation of a percutanious device was performed in the areas of bone interaction and fixation and neural interface control. A theoretical implant was designed for a 90^th percentile male, using a HA-G-Ti composite material to provide a permanent base to which the hand could attach. Using a radial implant length of 1.87 inches and an ulna implant length of 1.32 inches, the simulated implant could withstand a push out force of 10.260 pounds. Using nerve guidance channels and microelectrode arrays, a Regenerative Neural Interface was postulated to control the implant. The use of Laminin-5 was suggested as a method of preventing the lack of wound closure observed in percutanious devices. The exterior portion of a permanent artificial hand was analyzed by the construction of a robotic hand optimized for weight, size, grip force and wrist torque, power consumption and range of motion. Using a novel dual drive system, each finger was equipped with both joint position servos as well as a tendon. Fine grip shape was formed using the servos, while the tendon was pulled taunt when grasping an object. Control of the prosthetic hand was performed using a distributed network of micro-controllers. Each finger’s behavior was governed by a master/slave system where input from a control glove was processed by a master controller with joint servo and tendon instructions passed to lower-level controllers for management of hand actuators. The final weight of the prototype was 3.85 pounds and was approximately 25% larger than the 90th percentile male hand it was based on. Grip force was between 1.25 and 2 pounds per finger, depending on amount of finger flexion with a wrist lifting torque of 1.2 pounds at the center of the palm. The device had an average current draw of 3 amps in both normal operation and tight grasping. Range of motion was similar to that of the human model. Overall feasibility of the device is examined and factors involved in industrial implementation are discussed.