UTS RI Talk by Marco Sewtz

Speaker: Marco Sewtz - Institute for Robotics and Mechatronics, The German Aerospace Center (DLR)

Bio: Marco Sewtz received his B.Eng. degree in electrical engineering at the University of Applied Sciences of Munich and his M.Sc. degree at the Technical University of Munich. He works at the Institute for Robotics and Mechatronics at the German Aerospace Center (DLR) as a researcher since 2018. His interests focuses on SLAM and multi-model perception of the environment. Before his current role, he worked as an electrical designer for high-performance processing modules for space hardware at Airbus Defence and Space.

Abstract: In the realm of SLAM system development, datasets and technical metrics play a vital role in minimizing pose estimation error and frontend drift. However, integrating SLAM into robotic platforms poses challenges due to the platform's original design or limited computational resources. Additionally, SLAM systems must be adaptable to utilize pre-recorded maps and accommodate real-world changes, ensuring robust performance during development and operation.

This presentation will explore how SLAM system requirements are derived from real-world applications, highlighting practical considerations that shape design choices. Insights gained from integrating SLAM on different robotic platforms will be shared, focusing on lessons learned and strategies to achieve seamless modularity within the overall architecture. Moreover, we will discuss the demand for complete maps in specific applications and how adaptable SLAM frameworks enable the utilization of pre-recorded maps while effectively addressing environmental changes. Lastly, we will touch upon the current and future development of SLAM technology at the German Aerospace Center (DLR), offering a glimpse into the ongoing research and innovations.