No abstract

We consider the design and control problem of concentric tubes used in stereotactic neurosurgery. The goal is to optimally reach a configuration of the cannula linking an entry point on the skullcap to a pre-specified region inside the brain. Key issues related to this task are the mechanical behaviour of the cannula and the topography of the brain. We formulate an optimal control problem in order to determine a feasible path while minimizing brain damage caused by missing follow-the-leader behaviour. Numerical results show the potential of the proposed approach.

This paper deals with the special challenges in modelling and controlling fast-switching solenoid valves with Modelica. For this, a solenoid injector for application in direct injection combustion engines with switching times around 100 µs is being used as an example. The occurring eddy current losses as well as the local saturation phenomena based on magnetic field displacement require for detailed network models of the magnetic domain. Therefore, a new Modelica magnetics library based on a consistent systems modelling analogy is being proposed and implemented, which increases solvability of the injector models compared to the Modelica Standard Library. Additionally, different one-dimensional contact mechanic models for representation of the bouncing behaviour of switching-type solenoids are evaluated. The complete electro-magnetomechanical model of the injector is then used for synthesis of novel closed-loop control schemes found by model inversion and parameter optimisation.