The 'reciprocity' and 'consistency' based approaches to uncertainty identification for compliant motions

Two different approaches (reciprocity and consistency) to deriving identification equations for an arbitrary set of model uncertainties, and for arbitrary motion constraints are compared. For elementary motion constraints, the approaches are proved to be equivalent. For composite motion constraints, the identification equations are easily derived by the twist-based reciprocity approach and the wrench-based consistency approach. If the derivation of the identification equations is performed in the feature frame, the results are very simple and intuitive, as well as sufficiently efficient for online applications.<
>