Optimal trajectory planning for robots under the consideration of stochastic parameters and disturbances

Efficient control strategies of robots should cause only low on-line correction expenses. Hence, the mostly available statistical and a priori informations about the random parameters and disturbances of the underlying mechanical system and its environment should be considered already for off-line programming of robots. Measuring the violations of the basic mechanical conditions by means of expected penalty costs, a stochastic optimization problem is obtained for the computation of an optimal open-loop control. The stochastic optimization problem can be solved — after discretization — by parameter optimization.