Papers
13
Total Citations
177
H-Index
7
About
Thomas Beckers is a leading researcher at the intersection of machine learning, control theory, and robotics, whose work focuses on data-driven control of complex dynamical systems. His core contributions lie in leveraging Gaussian processes (GPs) to enable model-based control where traditional physics-based modeling is infeasible—such as in soft robotics and human-robot interaction. His highly cited 2017 paper on feedback linearization using GPs (68 citations) pioneered a Bayesian nonparametric approach to learning system dynamics, while his 2019 work on stable model-based control for robot manipulators (27 citations) demonstrated how GP regression can achieve high-performance computed-torque control without precise prior models. Beckers has also made significant advances in safe learning, introducing "smart forgetting" for online learning (11 citations) and developing stable tracking control for Lagrangian systems (18 citations). His recent tutorial on safe physics-informed machine learning (2025) provides a comprehensive framework for integrating physical knowledge with safety guarantees. With over 170 total citations, Beckers’ research is shaping the future of autonomous systems, enabling robots to learn and adapt safely in real-world environments.
Research Focus
Key Achievements
Top Papers
- 1Feedback linearization using Gaussian processes68 citations · 2017
- 2
- 3Prediction With Approximated Gaussian Process Dynamical Models25 citations · 2021
- 4Stable Gaussian Process based Tracking Control of Lagrangian Systems18 citations · 2018
- 5Smart Forgetting for Safe Online Learning with Gaussian Processes11 citations · 2020
- 6
- 7Safe Physics-informed Machine Learning for Dynamics and Control7 citations · 2025
- 8
- 9
- 10