Control of a multifunctional hand prosthesis

Barkhordar, Mohammad (1988) Control of a multifunctional hand prosthesis. University of Southampton, Doctoral Thesis.

Abstract

Work at Southampton University has resulted in a control strategy that utilises touch, slip and position feedback information to minimise the conscious effort required to control a multifunctional hand. Based on this principle, a fully articulated multidegree of freedom hand has been developed which closely replicates its human counterpart in features such as weight, shape and size. This thesis describes the work carried out to prepare the hand for everyday use by the limb deficient population. This involves the development of transducers suitable for use with cosmetic gloves and real-time implementation of the Southampton strategy on a portable electronic controller. The past work in the `below elbow' prosthetic field is reviewed with an intention to highlight the features of the existing devices. A detailed evaluation of the prototype hand is then carried out and a development environment in the form of a minicomputer set up to provide the facilities needed to undertake the required work. With an intention to produce a multi-purpose transducer compatible for use with cosmetic gloves, a range of touch, slip and force transducers are developed. A unique optical touch transducer is described which relies on a pocket of air in the glove material to resolve a signal that varies in proportion to touch force. Based on this principle, also described are a slip and force transducer constructed using a subminiature microphone and a pressure sensor respectively. Finally, the suitability of strain gauges is considered as touch and force transducers. A novel strain gauge digitizer is devised to simplify their interface to the controller. The control strategy is initially implemented on the minicomputer where the control algorithms are perfected to optimise the hand performance. Later on a dedicated microprocessor implementation resulting in a portable controller is presented. In addition to hardwired smoothing of the EMG supervisory signal, a further reduction in the controller size is demonstrated by performing this function in software. Suggestions are made for future work based on the experience gained during this project.

This record has no associated files available for download.

Contributors

Download statistics