Time-optimal control of articulated mechanisms

摘要

The (numerical) solution of the time-optimal control problem of articulated mechanisms (particularly of robotic manipulators) is considered. It involves an approximation of the prior control set by means of a strictly convex set, with any accuracy in the Hausdorff sence. This approach produces continuous optimal controls which may be used to provide nominal inputs for on-line control of articulated systems. Furthermore, the optimal controls thus obtained are shown to converge (in a weak topology) to a bang-bang solution if the error of approximation approaches 0. The application of the algorithm proposed here to the time-optimal control of a two-link manipulator is presented as a numerical example.