A robust grid adaptive optimization method applied to robot trajectory training

This paper reports a new approach to the multiparametric grid adaptive optimization method based on robust cost function applied to the trajectory training of a robot. The optimization scheme is based on using the calculation of a robust cost function around a central point vector in the space of parameters. A new central point vector point is chosen for that point which gives the smallest cost function and the calculation follows. The iterative calculation is stopped when the cost function reaches a given value. The trajectory has been simulated for a 3 degree of freedom robot described by the classical kinematical equations