Optimal solutions to the robot neighboring inverse kinematics

An optimal control approach to the problem of neighboring inverse kinematics is proposed for improving the accuracy of a manipulator. Two cases are studied. The first involves correction after kinematic parameter identification, assuming that the kinematic parameter errors are provided as the result of robot calibration. The other involves correction without a kinematic parameter identification phase, assuming that the world-coordinate poses of the manipulator are measured online. Based on linearized error models, the additive corrections of joint commands are found by a linear quadratic algorithm without explicitly solving the inverse problem for the actual robot. The algorithm is particularly suitable for real-time applications as shown by simulations