Efficient implementation of iterative inverse kinematics in real time control architecture

The paper proposes a procedure for the iterative computation of a robot inverse kinematics that uses different algorithms in synergy, so to obtain a singularity-robust and rapidly convergent technique. The application of such a procedure allows the creation of a unified version of the executable code for large classes of industrial robots, with the possibility to introduce new kinematic chains or more complex structures without the need to entirely rewrite the code. The paper offers also some guidelines about the software implementation and optimization in the industrial environment, and it is completed by experimental results where the procedure is applied to real industrial robots.