In this paper, a simple and efficient formulation is presented to optimize the kinematic error mapping from the Cartesian space of the end-effector position and orientation to the inputs space. The results are optimized allocation of the joints’ errors for the accepted end-effector positioning errors. The linearized relation (Jacobian) between the end-effector position and the joints’ errors is used. Upper bounds of the error are developed so that the error function is defined in the form of posinomials of joint errors. An optimization problem is set up to find the optimum error allocation at each joint. The problem is then formulated as a zero degree-of-difficulty geometric programming. It is shown that the weight of each term in the total cost function is constant and independent of manipulator’s design. As a result, the solution to the error optimization is readily available. Explicit solutions to these weight factors are presented.

This content is only available via PDF.
You do not currently have access to this content.