Modeling and Estimation of the Pitch Angle for a Levitating Cart in a UAV Magnetic Catapult Under Stationary Conditions

The paper presents a method for modeling and estimating the orientation of a launch cart in the magnetic suspension system of an innovative UAV catapult. The catapult consists of stationary tracks lined with neodymium magnets, generating a trough-shaped magnetic field. The cart levitates above the tracks, supported by four containers housing high-temperature YBCO superconductors cooled with liquid nitrogen. The Meissner effect, characterized by the expulsion of magnetic fields from superconductors, ensures stable hovering of the cart. The main research challenge was to determine the cart’s orientation relative to the tracks, with a focus on the pitch angle, which is critical for collision-free operation and system efficiency. A dedicated measurement stand equipped with Hall sensors and Time-of-Flight (ToF) distance sensors was developed. Hall sensors mounted on the cart’s supports captured magnetic field data at specific points. To model the tracks, the CRISP-DM (Cross Industry Standard Process for Data Mining) methodology was employed—a structured framework consisting of six stages; from problem understanding and data preparation to model evaluation and deployment. This approach guided the analysis of data-driven models and facilitated accurate pitch angle estimation. Evaluation metrics, including mean squared error (MSE), were used to identify and select the optimal models. The final model achieved an MSE of 0.084°, demonstrating its effectiveness for precise orientation control.