Kastor Felsner

Profactor (Austria), Institute for High Performance Computing and Networking

Papers

Total Citations

H-Index

About

Kastor Felsner is a researcher advancing the intersection of robotics and nondestructive evaluation, with key contributions in robotic coverage path planning and simulation-based machine learning for ultrasonic inspection. His work addresses critical challenges in automated quality control, particularly for inspecting complex, free-form 3D shapes using industrial robots. His most cited paper (2021, 5 citations) introduces a novel method for planning robotic motion to perform ultrasonic inspection on arbitrary geometries, a significant step toward automating inspection in industries like aerospace and manufacturing. Building on this, his 2022 paper (3 citations) tackles the persistent data scarcity problem in machine learning for nondestructive testing by proposing a domain randomization approach that trains neural networks exclusively on simulated data. This work enables scalable, cost-effective generation of training datasets without relying on expensive real-world measurements. Felsner’s research is notable for its practical focus on bridging simulation and reality, making automated inspection more accessible and reliable. His contributions are particularly valuable for students and researchers exploring robotic automation, computer vision, and deep learning in industrial applications, demonstrating how simulation can overcome real-world data limitations.