User:Thomas Denninger/sandbox

Prosthetic Hands

A prosthetic hand is a device that substitutes for a missing hand. Some prosthetics are merely cosmetic while others are functional and help people complete day to day tasks. This article focuses on prosthetic hands for transradial amputees. A transradial amputation is an amputation performed below the elbow. An ideal prosthetic hand is highly dexterous, controllable, provides sensory feedback, and is aesthetically pleasing. Modern prosthetic hands have great dexterity and aesthetics but control and sensory feedback are still limited. It is also very important that a prosthetic is instinctive to control so the patient accepts it as a part of their body.

Currently the best control systems available commercially use myoelectrics. Electrodes placed on the patient’s arm measure muscle activations and use these signals to control the prosthetic. There are several advantages of using myoelectrics; this system is un-invasive and relatively cheap. Some disadvantages include the lack of sensory feedback and limited bandwidth. Researchers are currently looking at neuroprosthetics as a new control system in which electrodes are surgically connected to the patient’s nerves. Signals from the nerves can be processed and turned into commands to move the prosthetic. This system could also provide sensory feedback by simulating nerve signals that would be generated by an intact hand.

1. Control and Sensory Feedback

1.1 Myoelectrics

Myoelectrics or electromyography is a very popular choice for controlling prosthetics. To control prosthetic hands electrodes are placed on the skin of the patient’s forearm. The electrodes measure muscle activation potentials in the patients arm. These signals indicate which muscles the patient is willing to activate and with what force. One study found that after a patient has been familiarized with the use of the myoelectric controlled prosthetic, the system can correctly perform the type of grasp the patient wants with an accuracy of 87.67%. [1]

There are two main limitations with myoelectrics. Myoelectrics have a limited bandwidth so it can only control a couple of degrees of freedom. [2] People have forearms of various strengths, shapes, and sizes. The muscles in the forearm are not just involved in the motion of the hand they also control the arm. All of these variables decrease the accuracy and consistency of myoelectrics. Also some people have too much of their arm amputated and they do not have enough remaining muscle to control the device. Muscle atrophy is another problem, atrophied muscles do not produce strong enough electric signals to control the device. [6]

Among the many advantages of myolecectrics is that it is non-invasive; electrodes are placed on the surface of the skin. Simpler myoelectric systems are very affordable while complicated ones that control several degrees of freedom are more expensive. Most amputees live in developing and third world countries [6]. Myoelectrics is rather inexpensive and is accessible financially for people in these poorer countries.

One of myoelectrics’ greatest shortcomings is its lack of sensory feedback. [2] Researchers have placed sensors on the fingertips of a robotic hand that detect when they are in contact with an object. These sensors then trigger servo motors that apply pressure on the skin of the end of the amputee’s arm. The pressure applied by the servos is proportional to the grasping force measured on the robotic hand. These servos are placed in a pattern corresponding to the fingertips; this provides a similar but displaced sensation on the amputee’s skin. This technology however is not yet available commercially. [2]

1.2 Neuroprosthetics

Neuroprosthetics directly read and interprets neural signals as well as produce artificial neural signals. Tactile feedback from a sensorized prosthetic is achieved by characterizing neural signals that are produced in an intact limb and simulating them through electrical stimulation of the residual peripheral nerves. [3] There are several types of receptors in the hands that send various neural signals. All of these signals must be analyzed so a computer model can reproduce them based on inputs from the prosthetic hand. [3] Neuroprosthetic Diagram [4] Signals from the amputee’s brain can be read from the efferent nerve fibers and translated into commands to control a prosthetic hand. The control algorithms sort the spikes and noise and generate the commands. Testing has been performed on animals and the results are very promising. [4] A 26 year old human male had his arm amputated after a car accident. He had several electrodes implanted in the remaining part of his arm for four weeks. There weren’t any complications from the procedure and the subject was able to control three different grasps with over 85% accuracy. [4]

Neuroprosthetics have great potential but are not commercially available yet. Neuroprosthetics use the body’s nervous system thus the control and feedback mechanisms are very natural. Neuroprosthetics will give an amputee unprecedented control and sensory feedback along with a greater feeling of ownership. Another advantage of neuroprosthetics is the bandwidth, more information can be transmitted using neuroprosthetics than electromyography. [4] This will allow the user to control more degrees of freedom. The greatest disadvantage of neuroprosthetics is that it is invasive. Surgery is required to attach the electrodes to the nerves in the patients arm, however as neuroprosthetics advance it will get less invasive.

1.3 Shear Sensors

In order to provide a sense of touch, sensors are needed on prosthetic hands. Basic touch sensors can detect when they are in contact with an object and how much pressure is being applied. Researchers have recently developed a new sensor to detect shear forces which are lateral forces applied parallel to the surface of the hand. The shear sensors will be able to detect if an object is slipping or deforming. This is crucial when grabbing delicate and oddly shaped objects such as eggs. [5] Accuracy, repeatability, and cost are very important for tactile sensors. This particular sensor is designed for forces up to 4 Newtons, is accurate, and repeatable. These shear sensors only cost $2 when made in small quantities. If these sensors are mass produced it would drive the cost down. [5]