Electromagnetic near-Field enhanced UAV...

The overall goal of the project EMFLanding is to lay the foundation for continuous operation of UAVs in harsh outdoor environments by establishing a novel electromagnetic near-field (EMF)-based sensor technology including an estimation framework that correctly fuses sensor information with very different characteristics for accurate UAV state estimation. The primary goal is comprised of the following sub-goals:• Design of an electromagnetic near-field-based sensor system for accurate UAV state estimation:In a first step, the physical field emitters and inductive pick-up coils of both sending and receiving unit will be optimized to cover a suitable range with sufficient resolution while respecting size and power consumption restrictions. The influence of electromagnetic interference (EMI) caused by the electric drives of UAVs will be analyzed and minimized by finding suitable lock-in type demodulation techniques and spreadspectrum approaches.• Real-time forward and inverse solutions of the underlying electromagnetic-near field equations:As accurate, real-time information about the relative position of the UAV with respect to the landing platform is crucial for safe landing, we will investigate novel, physics informed neural networks in combination with regular neural networks for fast inference in addition to a classical model-based approach. • Development of a multi-sensor fusion framework to accurately determine the state and associated uncertainty of the UAV in real-time:The novel measurement principle of EMF-based localization will be included in a state-of-the-art multi-sensor fusion framework. An important element is to accurately quantify the associated uncertainty of the neural network prediction to allow for consistent state estimation and seamless switching between global position sensors and EMF-based localization. • Proof-of-concept of the landing scenario:We will illustrate the suitability of the EMF-based localization during landing through proof-of-concept experiments.