Optimisation of anthropomorphic robot hand design through human manual dexterity testing

Abstract

In this work, a new systematic approach towards the classification of the attributes of anthropomorphic hands with respect to their influence on manual dexterity is proposed. This approach involves the application of standard manual dexterity tests, normally applied in medically related fields and in industry, on healthy human subjects who have specific attributes of their hands selectively constrained or blocked for the duration of the tests. The contribution of the constrained attributes towards manual dexterity can then be assessed from the test results. To demonstrate this method, we have investigated the contribution of different degrees-of-freedom of the human hand to its performance rating in the dexterity tests, with the objective of using the results to help us design a new robot hand that maximizes dexterity through optimal use of resources. Further applications of this approach are also discussed.