Postdoctoral Researcher in Verification & Validation of learning-based perception systems for robotic proximity operations

The successful candidate is expected to take a leading role in defining, acquiring, managing, and scientifically contributing to projects around AI-enabled space-borne perception systems for robotic proximity operations in collaboration with Redwire Space Luxembourg. The candidate will carry a leading role in this area and support PhD candidates in their thesis research. The candidate will work closely with Prof. Olivares-Mendez and Dr. Carol Martinez, the members of the Space Robotics (SpaceR) research group (www.spacer.lu ) and Redwire Space Luxembourg (https://redwirespace.com/ ).

The group works across the entire spectrum of space robotics research and applied research to conceive and demonstrate future orbital and planetary space robotics paradigms on in-space servicing, assembly, and manufacturing, space debris removal, XR immersive teleoperation, robot multi-modal perception (vision and tactile), and multi-robot cooperation. Researchers with an interest in non-terrestrial robotic manipulation will find a young and vibrant team of over 23 members fostering a collaborative atmosphere (1 Prof., 1 Research Scientist, 15 PhD students, 5 PostDocs, 1 Research assistant). SpaceR’s state-of-the-art facilities include two physical laboratories, the LunaLab and Zero-G Lab, and a new Robotic Manipulation Lab – dedicated environments for researchers to test their results on advanced equipment. For further information, you may check: wwwen.uni.lu/snt/research/spacer and www.spacer.lu

The candidate will lead the development of a V&V framework for AI-augmented perception systems and will be responsible for activities in the Zero-G Lab and contributing to the Black-Hole Lab at Redwire Space Luxembourg for Hardware in the loop emulation of on-orbit scenarios.

Responsabilities:

• Investigate existing verification and validation (V&V) techniques for space systems, software and algorithms with a focus on specific challenges of space-borne perception and proximity operations uncooperative spacecraft .
• Develop novel methods for characterization and quantification of input-output bounds in an AI-augmented perception system: Develop novel methodologies and tools for characterizing and/or quantifying the performance envelope, robustness, and safety properties of perception systems that encompass learning-based models.
• Develop methods and tools for automatic/procedural generation of relevant operational conditions
• Inform and educate certification processes for such systems for the industrial partner: Explore approaches to generate safety cases and evidence suitable for contributing to the potential certification of these AI components according to relevant space standards.
• Lead SITL and HITL Test campaigns: Implement algorithms and conduct experiments using both synthetic and potentially real datasets/simulators relevant to space environments.
• Teamworking: Collaborate closely with team members at SpaceR and Redwire Space