Real-time dynamic model identification for a new generation of cobots

In this research you will contribute to the development, implementation and experimental validation of the novel control framework. This framework will take into account the payload dependent dynamic model of the robot system, including its flexibilities, in order to avoid or compensate vibrations resulting from the latter. Your focus will be on the development, implementation and experimentally validation of new off-line and on-line data-driven modelling techniques: optimization based combined model structure selection, parameter and state estimation. You will research regularized estimation to find trade-offs between model complexity and accuracy, estimate model uncertainty to account for in the robust control design. You will first work on numerical models of the novel cobot, and gradually, as the hardware becomes available move towards experimental application and validation. You will closely cooperate with the other PhD researchers involved in this project and that focus on the mechanical design and controller development.