131.
techniques. They do not -require the computation of the manipulator Jacobian or its inverse nor do they require computation of the forward kinematics at any time before convergence is achieved. When the initial estimate of the variable is reasonably close to a solution (typically within 5, degrees) -convergence is- achieved within six ,iterations even with accuracies better than 10O5 which makes real-time
high precision inverse kinematics a reality. Finally, 'the
algorithms simplicity is such that they can be programmed and executed very rapidly even on a personal microcomputer for any manipulator with at most six degrees of freedom. This programming simplicity makes the one- and twodimensional iterative techniques suitable for use as offline interactive inverse kinematic tools as well.
The disadvantages of the inverse kinematics methods
developed here are those inherent to iterative algorithms. No theoretical guarantee of convergence can be given and the number of iterations required for convergence depends highly on the initial estimate of the variables. A disadvantage
over the homotopy map method (Tsai and Morgan 1984) is that convergence to only one solution is possible, again a common
disadvantage of iterative techniques. Searching for more
than one solution requires trying various estimates of the variables.
One problem that the methods developed in this text share with Tsai and Morgan's homotopy map method (1984) is
..