Copy the page URI to the clipboard
Kyberd, Peter and Poulton, Adrian
(2012).
Abstract
Introduction
Humans control their body by integrating different sensory inputs to create knowledge of the body’s orientation and disposition. This allows them to react to changing circumstances with little apparent cognitive load. When a person is using a prosthetic limb, some of these modalities may be absent, but many more are hard to integrate into the control of their prosthesis. An example is the ability to maintain the hand orientation while the body position changes as the person gets up from a chair or reaches forward. This is simple with an intact arm, but when there is a loss above the elbow, the ability to correct the change in angle of the forearm as the arm reaches forward, is challenging. It would become even more difficult if there was some humeral deviation at the same time, requiring wrist pronation as well. If additional sensors are added to the arm then this task can be devolved to the electronic controller.
Method
The second generation ToMPAW arm (and updated device based on the previous design), contains microcontrollers for the elbow, wrist and hand, these can take additional information about the angle of the different joints, as well as integrate additional inputs from accelerometers in the arm. While this idea was put forward as part of the Southampton Arm project in the 1980s, it is only with the advances in controllers and cheap compact accelerometers developed for consumer goods, that these ideas can be pursued. The hand, elbow and wrist are controlled sequentially, switched from the shoulder harness. The hand is controlled as a Southampton Hand, so when it is instructed to hold an object, the rest of the arm is informed of this state and only under these circumstances will the orientation of the wrist and forearm be maintained. If the user switches control to the elbow or wrist they can override the position and change it. Once control is switched back to the hand the automatic control resumes. It is turned off completely when the hand has released the object.
Results
The arm has been prepared for the single subject who was a long term user of the first generation arm. The velocity of the correction motions has been reduced so that the arm does not over react to small changes in attitude, but steady co-ordinated movements can be corrected for, until the wrist or elbow reaches their range limits.
Discussion
This is one application of the additional information that can be garnered from additional sensors. The critical factor for acceptance of any automatic feature is the ability of the user to predict when it will be engaged and what its action will be. By linking the corrections to when the arm is holding objects the user can be sure of when the automatic action will be engaged and when they can turn it off.
Conclusion
Microprocessor controllers and cheap sensor technology can be combined to give prosthetic arms more autonomy.