Life Sciences

Model reduction for predictive online MR-Thermometry

Predictive, model-based simulations can help treating physicians by offering real-time assistance, for example by predicting the temperature profile in the liver during a minimally invasive thermal ablation for the treatment of liver tumors. Such a simulation requires the continuous solution of an accompanying patient-specific parameter identification problem during the operation. The online optimization requires a substantial reduction of the model complexity. A combination of model order reduction techniques and space-mapping strategies will be used to achieve this.

Contact:  Kevin Meligan

Control of Glucose Balance

Model- and simulation-based medical decision support gains increasing importance in health care systems worldwide. It allows predictions and optimizations for individual patients. This research focuses on modeling and optimal control of the glucose balance of patients in intensive care units. The existing bio- medical models are compartment models of differential equations that describe the temporal evolution of the blood glucose and the insulin concentration in the human body, but neglect crucial effects and relevant stochastic influences. Collaboration with Center for Model-based Medical Decision Support (MMDS), Aalborg University, Denmark.