Mats Steinweg

Robotics Research (United States), RWTH Aachen University

Papers

Total Citations

214

H-Index

About

Mats Steinweg is a robotics and machine learning researcher whose work spans autonomous aerial systems, deep reinforcement learning, and sensor-based perception. He is best known for his influential 2021 paper "Autonomous Drone Racing with Deep Reinforcement Learning," which has accumulated over 185 citations and stands as a landmark contribution to the field of agile autonomous flight. This work tackled one of the core challenges in drone racing — planning time-optimal trajectories without assuming perfect prior knowledge of waypoints — by leveraging deep reinforcement learning to enable drones to navigate complex courses at high speed with remarkable adaptability. The paper's significant citation impact reflects how it has shaped subsequent research in both autonomous navigation and competitive robotics. Beyond aerial systems, Steinweg has contributed to robust perception pipelines, notably through his work on self-supervised person detection using 2D range data and camera calibration. This research addresses critical limitations in LiDAR-based detection, particularly the scarcity of annotated training datasets, proposing self-supervised approaches that improve generalizability across environments and sensor configurations. Together, these contributions position Steinweg as a thoughtful researcher bridging the gap between theoretical machine learning and real-world robotic deployment, with demonstrable influence across both communities.