Trajectory optimization for kinematically redundant arms

Abstract

Abstract A review of local optimization methods for resolving joint configurations in underconstrained manipulation tasks is conducted. A new approach is developed for observing joint limits and avoiding obstacles during the trajectory planning. The methodology is used in a four‐link arm example to avoid a workspace singularity and is compared with results using the extended Moore‐Penrose technique. An alternative measure of arm “manipulability” based directly on the rank of the Jacobian is also introduced.