We demonstrate a method for self-sensing of a magnetically actuated prism that can be used, e.g., in a feedback-loop without the need of additional sensors. In order to use the impedance of the actuation coils as a measurement parameter, we first obtained the optimal measurement frequency that is well separated from the actuation frequencies and at the same time provides the best compromise between sensitivity to the position and robustness. We then developed a combined actuation and measurement driver, and correlated its output signal to the mechanical state of the prism using a defined calibration sequence. We demonstrate that we can reliably measure the state of each actuator and determine the tilt angle of the prism with an accuracy of $\pm 0.1$${}^{\circ }$ in the polar angle over a range of $\pm 4$${}^{\circ }$ and $\pm 20$ mrad in the azimuthal angle.

Electromagnetic actuators are widely used in various application areas, from electrical machines to microsystems. A popular design goal is the development of multistable actuator concepts that enable energy-efficient switching between stable positions. Typically, the finite element method (FEM) is used for optimization, thus minimizing the need for extensive experimental trials. A novel design for a multistable magnetic actuator tailors stable positions using mu-metal rings. This paper details the calibration of a numerical model based on experimental data, as well as a design exploration to examine the impact of ring dimensions on performance characteristics. The results demonstrate that specific actuator requirements can be directly influenced through appropriate variation of ring parameters. Additionally, a simplified modeling approach is prepared, where the force-displacement curve of an actuator with multiple stable positions is approximated by superimposing the individual components' curves. This approach allows for a highly efficient preliminary design optimization compared to a FEM-exclusive approach. Candidate designs can be refined further in a more comprehensive model or experimentally validated.