MOTION PLANNING FOR PDEs - A FLATNESS-BASED APPROACH
Séance du jeudi 17 Décembre 2015, Salle L 111, 16h30
Thomas MEURER, Prof. Dr.-Ing. habil., Chair of Automatic Control, Christian-Albrechts-University, Kiel, Germany
Mathematical modeling of a dynamical process leads to a distributed parameter description in terms of partial differential equations (PDEs) whenever spatial and temporal effects have to be taken into account to properly represent the system dynamics. Well known examples comprise fixed–bed reactors in chemical engineering, reheating processes in steel industry, elastomechanic smart structures with embedded actuators and sensors as well as fluid flow and fluid–structure interactions. Moreover, novel applications emerge such as modeling and control of interconnected multi–agent systems or robotic swarms, that allow for a continuum description by means of PDEs. In addition to the feedback stabilization problem, motion planning and tracking control to impose a desired transient behavior as well as distributed parameter state observers for the realization of model–based control concepts and for process monitoring have gained increasing interest both in control theory and applications. The presentation addresses recent developments for motion planning and tracking control given PDE systems. This includes in particular flatness–based techniques for motion planning and feedforward control design exploiting, e.g., spectral properties of the system operator or formal integration of the PDE. In addition, it is shown that the suitable combination of motion planning with backstepping or passivity–based feedback control concepts allows to systematically determine stabilizing tracking controllers. These enable to accurately and robustly realize the desired spatial–temporal evolution of the state variables. Complementing simulation results for problems arising in different physical and technical domains the applicability of the developed design techniques is illustrated by means of experimental data obtained for coupled beam structures with embedded piezoelectric actuators.
Thomas MEURER, Prof. Dr.-Ing. habil., Chair of Automatic Control, Christian-Albrechts-University, Kiel, Germany
Mathematical modeling of a dynamical process leads to a distributed parameter description in terms of partial differential equations (PDEs) whenever spatial and temporal effects have to be taken into account to properly represent the system dynamics. Well known examples comprise fixed–bed reactors in chemical engineering, reheating processes in steel industry, elastomechanic smart structures with embedded actuators and sensors as well as fluid flow and fluid–structure interactions. Moreover, novel applications emerge such as modeling and control of interconnected multi–agent systems or robotic swarms, that allow for a continuum description by means of PDEs. In addition to the feedback stabilization problem, motion planning and tracking control to impose a desired transient behavior as well as distributed parameter state observers for the realization of model–based control concepts and for process monitoring have gained increasing interest both in control theory and applications. The presentation addresses recent developments for motion planning and tracking control given PDE systems. This includes in particular flatness–based techniques for motion planning and feedforward control design exploiting, e.g., spectral properties of the system operator or formal integration of the PDE. In addition, it is shown that the suitable combination of motion planning with backstepping or passivity–based feedback control concepts allows to systematically determine stabilizing tracking controllers. These enable to accurately and robustly realize the desired spatial–temporal evolution of the state variables. Complementing simulation results for problems arising in different physical and technical domains the applicability of the developed design techniques is illustrated by means of experimental data obtained for coupled beam structures with embedded piezoelectric actuators.