Prosthetic Hand Control Using EMG Signals

An amputee’s ability to control a prosthetic limb is greatly hampered if the limit function of data transmission does not match the required transfer function of data reception. Robust and intuitive control of prosthetic hands is currently provided by remote control systems that utilize teleoperated master-slave configurations. Such systems are just as much a limitation to an amputee’s freedom as they are an important step towards the integration of neuroscience and robotics. To replace a lost hand or to control a prosthetic hand intuitively, the prosthetic hand must follow the same input signals that a natural hand can understand. To create a prosthetic limb that mimics a human hand, the input signals that would yield analogous motion must be identified. Natural hand movement is a result of intricate, parallel neural processes alongside an exceptionally vast network of multiplexed control signals reaching forearm muscle motor units. Functional hand movement results from the grouping and varying of synergistic finger actuator signals at the minutiae of the unit level. The forearm is inhabitable for any type of invasive robotic mechanism. A hand prosthesis requires a compact and low-cost alternative to the embedded actuators and sensors that crowd the forearm. Understanding the anatomy of the forearm is crucial to a non-invasive, compact, and low-cost system.

The action of group finger motion frequency is embedded in the finger’s fundamental natural movement. Fractionature further groups the buttons to uninhibited motion. The robust mechanical behavior of this granular action provides varying level control throughout a large motion range. Understanding the properties of signal redundancy can yield a compact, low-cost sensor network. Utilizing low-cost sensor alternatives to develop a rapid, low-resolution grasp detector can yield component-level behavior patterns. Coarse-to-fine control is the opposite of traditional approaches; one would first control the global motion and then move down to the minutiae. Simulating finger motion is very useful for designing the prosthetic hand and prior experimentation. With basic control and motions simulated, control algorithms could be verified once the sensors were placed on a physical body.