Model-free approaches for the energy minimization of robot trajectories

Conference contribution, 2014

Energy eciency is increasingly important for robotic manufacturing systems. An energy minimal schedule may be generated given that the energy consumption for each robot operation may be parameterized as a function of execution time. Existing approaches rely heavily on the use of models to generate functions describing energy consumption. We are interested in solving the problem using model-free approaches, such as Iterative Learning Control, model-free/iterative Dynamic Programming and derivative free optimization.