Comment on “automatic planning of manipulator transfer movements”

It is pointed out that because errors in a manipulator environment come from a variety of sources, a worst-case correction as described by Lozano-Perez (ibid., vol.SMC-11, no.10, p.681-9, 1981) is unduly restrictive, and could lead to the fallacious conclusion that certain movements are not possible, or to the overspecification of tolerances on system components. It is shown how a simple probabilistic model and some straightforward mathematics can be used to produce an expansion limit that is much less restrictive in path planning.