The Right Mix of Visual and Mechanical Constraints to Guide the Relative Pose between Rigid Objects

Restoring and/or obtaining the correct relative pose between rigid objects is a common precision task in many sectors, from industry to healthcare, as during the reduction of long-bones fractured. Augmented Reality (AR) can provide useful indications, showing how one bone should be positioned with respect to another, for example. However, moving a rigid body by manually managing its six degrees of freedom (DOF) until the desired pose is obtained can be difficult. At first, because AR visualization is commonly not able to provide an accurate depth perception. Second, because it is very challenging to control and handle all the six DOFs simultaneously. In this work, we evaluated whether the addition of adaptable and easily adjustable mechanical constraints, each constraining a single degree of freedom, allow increasing the feasibility, accuracy, and efficiency of an AR guidance. This preliminary study shows that handling simultaneously all the 6 DOFs of a rigid object is hardly feasible under hand- free AR guidance, and that we can therefore insert simple mechanical constraints to facilitate the achievement of the desired relative pose between rigid objects.