AUTOMATED INVERSE-KINEMATICS FOR ROBOT OFF-LINE PROGRAMMING

Automated methods are developed to classify a robot's kinematic type a nd select an appropriate library inverse-kinematic solution based on t his classification. These methods automatically generate Denavit-Harte nberg joint frame parameters, given any frame representation that can mathematically be represented as a homogeneous transformation. To redu ce the number of closed-form inverse-kinematics solutions required for a broad class of serial robots, additional methods account for differ ences in robot zero state, base frame location, and joint polarity. Fu rther generalization results from using joint frame decoupling to map lower degree-of-freedom robots into the inverse-kinematics solutions o f higher degree-of-freedom robots.