Karl Martin Kajak
Papers
1
Total Citations
6
H-Index
1
About
Karl Martin Kajak is a rising star in the field of autonomous space navigation, with a focused expertise in vision-based relative navigation for uncooperative spacecraft. His research directly addresses one of the most critical challenges in modern astrodynamics: enabling safe proximity operations around objects that provide no navigational aids, such as defunct satellites or space debris. Kajak’s major contribution lies in pioneering the use of domain randomisation combined with convolutional neural networks (CNNs) for keypoint-regressing pose initialisation. This innovative approach allows a monocular camera to robustly estimate the position and orientation of a target spacecraft—even when that target has a finite, symmetric shape—without requiring extensive real-world training data. His most-cited work (2023) demonstrates how this technique overcomes the "reality gap" in simulation-to-reality transfer, a key hurdle for practical on-orbit servicing and debris removal missions. While his citation count is still growing, Kajak’s work is already recognised as foundational for next-generation relative navigation systems, positioning him as a key contributor to the future of autonomous operations in space.
Research Focus
Key Achievements
Top Papers
- 1