IMPROVED CONTROL FOR AN ARTIFICIAL ARM

Abstract

Successful control of multi-degree of freedom upper limb prostheses generally uses some form of sequential instruction. This is because simultaneous control of multiple inputs requires a considerable concentration to be operated effectively. In contrast, the natural arm is controlled in a parallel manner with a high level of subconscious control. Such control uses feedback, the person is rarely conscious of the feedback information, and most of the control is automatic. Attempts to achieve similar control with a prosthesis would requires a wide bandwidth feedback channel to the controller. This is currently impractical if the controller of a multiple degree of freedom arm is the wearer, because only very low frequency feedback is achievable. The Southampton Arm control philosophy avoids this bottleneck by keeping the low level control within the prosthesis and leaves low bandwidth and strategic control to the operator [1]. This study describes the development of a control system for an artificial arm based on this concept. A joystick was used to input control signals via a harness at the shoulder, allowing a user to manipulate the arm with small movements of their acromium. A co-ordinate control strategy (inverse kinematics or IK) was implemented allowing both the shoulder and elbow joints of the arm to move simultaneously giving a smooth, more life like, motion. A trial was developed to asses the cognitive load required to operate the arm using the Dual Task paradigm. Tests were carried out on 5 subjects, using two different control strategies, IK and a direct control strategy (DC) using two degrees of freedom separately.