This book describes the technical design characteristics of the main components that go into forming an artificial hand, whether it is a simple design that does not have a natural appearance, or a more complicated design where there are multiple movements of the fingers and thumb. Mechanical components obviously form the structure of any hand, while there are some lesser known ideas that need to be explored such as how to process a slip signal. The focus of the book is the design of artificial hands for people, who through trauma or congenitally, only have one or no natural hands, with an emphasis on myolectric hands - powered hands that are controlled by the small electrical signals from residual muscles. An in-depth treatment of mechanisms, sensors, control, and hand assessment is included. Bringing together decades of research from the University of Southampton - a centre of excellence in this field - this book is essential reading for researchers and advanced students of robotics, prosthetics and mechatronics as well as professional engineers and prosthetists in universities, industry and hospitals who are involved in the design and manufacture of prosthetic hands.
Inspec keywords: mechatronics; design engineering; dexterous manipulators; sensors; prosthetics
Other keywords: sensors; prosthetic design; mechatronics; hand assessment; robot design
Subjects: Prosthetic and orthotic control systems; Biomechanics (mechanical engineering); General and management topics; Robot and manipulator mechanics; Design; Transducers and sensing devices; Manipulators; General topics in manufacturing and production engineering; Micromechanics (mechanical engineering)
Progress in the mechanical design, control and sensing for replacement artificial limbs has seen peaks in activity as the casualties of war return home from international conflicts. Simple mechanisms with open-loop control provide some gripping of objects. To form multiple grip patterns requires the addition of more actuators and the inclusion of sensors and closed-loop controllers. A major constraint is the mass of a hand as it is attached to the user's stump, and if it is too heavy, it will not be acceptable to a user.
A passive artificial hand that has no power source or control over its movement can be made from a range of materials whose physical properties are important to provide a lifelike visual impact, feel and durability. In contrast, a hand that needs to be designed to hold and manipulate objects requires as at least some form of mechanical structure, c.f., a skeleton but also other components so that it is functional. It has a palm or central structure for the attachment of finger joints and links that is similar to the anatomy of a natural hand. Also there is provision for actuators, i.e., electric motors or if the device is body powered, cables to react against the mechanism. For a powered anthropomorphic hand, choices are made about which of the fingers and thumb should be powered. A single actuator could move the first finger or groups of fingers and the thumb to form pinch grips where the fingertips oppose the thumb tip. With two actuators a finger or group of fingers can move independently of the powered thumb. Alternatively, one actuator could move both the first finger and thumb together and the second actuator moves the other fingers. Three actuators allow for the thumb to have two degrees of freedom with the third actuator powering all the fingers. Alternatively, the first finger could be powered with one actuator and the three other fingers with a second actuator. The powered thumb then has only one movement of extension/flexion and no rotation. The thumb can also have no power and be moved into position using a natural hand, in which case all three actuators are available for moving the fingers. With four actuators, there are more possible combinations for grip patterns but each additional actuator adds mass not only to the hand but also is in need of more coordination and algorithmic control of the fingers and thumb.
Signals from sensors are used in closed-loop feedback systems to automatically adjust variables such as grip force to arrest object slip. Also information from sensors can also be used to alert the wearer of a hand that an object is too hot to hold, for example. Resistive, capacitive, inductive, optical and Hall effect technologies can all be used to form sensors. Sensors can be made of low-cost and low-power technologies from screen printable piezoresistive and piezoelectric materials, which are used to measure forces in the fingers and thumb and to detect slip or surface texture. Analogue electronics in the form of instrumentation amplifiers condition the signal from a sensor. Digital processing of the analogue signal from a texture sensor mounted on a fingertip reveals the repeating pattern and coarseness of an object surface. A novel form of signal processing, using the rectified mean variance of the signal (equivalent to the FFT) is used in this analysis.
A person wearing an artificial hand will want to control the movement of the fingers and thumb or generate grip force. They will need some device that could be an electronic, electromechanical or mechanical interface. The person may simply desire to open and close the hand. An electronic signal generated from some action by the person can be sent to the actuator(s) to cause the desired effect. A simple open-loop controller can be made to achieve such an action (Figure 4.1). For the position of the fingers to be controlled precisely requires a detailed mathematical model of all the parts of the system. The model parameters will need to be determined through measurements made on the components. Only then can precise control be achieved. However, most systems will have variable parameters such as friction and backlash in the finger mechanisms. A much better idea is to use negative feedback shown in Figure 4.2. The advantages of using a feedback control system have been extensively researched. For the artificial hand, there are the advantages of improved performance in terms of making precise finger movements and automatic grips. The person controlling an artificial hand can control posture and force generated in an open-loop way, but this would require continuous monitoring of the hand and have a cognitive burden. Adopting a hierarchical control structure frees the person of having to think all the time about state of the hand (Figure 4.3). The lower levels of control are automatic electronic systems, so that the person simply instructs the controller to instigate grip of an object and monitor any slipping that may occur. People who use an artificial hand would like to have some sensory feedback, i.e., be able to feel some property of an object such as surface texture. Here, sensors can estimate these physical properties and through the skin surface stimulate neurons to elicit a sense of feeling, i.e., haptic feedback.
The assessment of natural or artificial hand function is wide reaching. Any procedure should satisfy a set of main criteria; reliability (inter-rater), standardised equipment, validated, portable and standardised metrics. All of these criteria are encompassed in the SHAP. The movement of an abstract test object or the manipulation of an everyday object used in ADL requires coordination of the neuromuscular system and can be related back to the time taken to carry out the task. The underlying physics and control of a moving object can provide evidence and give clues as to the way that the assessment is performed by a person. Making a smaller version of SHAP for the assessment of children requires careful consideration of materials especially for the design of the lightweight abstract test objects.
An outline of the engineering requirements for a prosthetic hand has been described in this book where the techniques and ideas presented are also applicable to robotic hands. Of all the components that form the system of an artificial hand, it is perhaps the algorithms, sensors and the processing of signals that will play an increasing role to improve the functionality of advanced hand designs. The new developments in 3-D printing (additive manufacturing) allow for custom-made designs, modular components and low cost in the development of designs. However, the polymer-based materials have non-isometric properties and can be relatively weak compared to blocks of materials that are machined into the desired shape. Also if a high tolerance is required, then a 3-D printed component may lack the precision required. This technology has seen considerable development in the recent years and opens up the possibility of low cost and modular component design. Further developments, especially using low-density metals, will greatly enhance the durability of future prosthetics.