Adaptive control of an arm prosthesis

Swain, Ian Douglas (1982) Adaptive control of an arm prosthesis. University of Southampton, Doctoral Thesis.

Abstract

This thesis is concerned with the control of a complete arm prosthesis. The principle on which this work is based is that the control should be as simple as possible for the wearer without loss of flexibility or resorting to a sequential system. To achieve this, the control of the human arm has been observed with particular attention being paid to those functions which are controlled by the lower levels of the central nervous system, i.e. below the level of conscious intervention. As a result of these observations an adaptive hierarchical control system has been proposed for a complete, eleven degree of freedom, hand/am prosthesis. This can be controlled in principle by only four inputs from the patient, three body movements and a ringle EPIC channel. The additional information demanded by the number of degrees of freedom is provided by constraints on the elbow position, analysis of wrist trajectory and by a variety of sensors on the hand and arm. A simplified version of this system including a six degree of freedom arm supporting a dummy hand with touch sensors, has been constructed. The control signals from the wearer are obtained from a harness built to fit the author. In the design of the prototype used to evaluate this simplified system consideration was given to the range of movemants of the arm, and to speed and torque requirements, but no attempt was made to produce a clinically acceptable prosthesis. The control system was implemented using a Digicon Micro 16v computer for which a Real Time language has been developed within the control group. The assessment of the performance of a prosthesis is inherently subjective, as the degree of control will vary from wearer to wearer. The keyboard in the form of set trajectories stored within the computer, enabled objective tests to be carried out. The results of which show that the arm performance was in accordance with the proposed theory. The results obtained using the harness to provide the input signals are more difficult to analyse as it is the overall pattern of movement that is important rather than the usual control performance criterion such as response time or steady state error. However, these tests show that the wearer can perform complex co-ordinated movements without undue conscious effort. Due to the difficulties in presenting information relating to three dimensional trajectories in either graphical or tabular form, the results of the test, using inputs from both the keyboard and the harness, are available on a video tape as supplement to the rather descriptive results given in this thesis.

This record has no associated files available for download.

Contributors

Download statistics