Energy-Based Modeling and Control

We develop and apply methods for modeling, discretization and control of dynamical systems under preservation – or inspired by – their physical structure. The port-Hamiltonian framework offers well-structured, energy-based models with clear power interfaces. We want to preserve this structure under spatial discretization and temporal integration to retain the modularity and physical consistency of the models, but also to solve control-related tasks (order reduction, feedforward and feedback control, optimization) in an adapted, beneficial manner. 

Current topics are sampled-data control based on various geometric integration techniques, e.g., Hermite interpolation, and (mixed) finite element (FE) approaches for models of nonlinear elastodynamics, which can be exploited for the systematic solution of various control tasks.

Another field of research are geometric methods for modeling and control of flexible mechanical systems that are based on global representations. This means that the (Lie group) manifold structure of the configuration is considered in the system models and the derived numerical and control schemes. This perspective allows for globally valid controls and the description of mechanical continua via exact beam models. With our flexible joint and flexible link modular lab manipulator, we want to showcase that such modern geometric approaches for modeling and control can be efficiently implemented on industrial standard PLC hard- and software.

Find a recent poster of our group here.