Jonas Ulmen

University of Kaiserslautern, University of Koblenz and Landau, Kempten University of Applied Sciences

Papers

Total Citations

H-Index

About

Jonas Ulmen is a researcher at the forefront of robotics, control systems, and machine learning, whose work bridges the gap between classical dynamics and modern AI. His primary research areas include dynamic parameter estimation, system identification, and the development of computationally efficient neural network controllers for resource-constrained platforms. Ulmen’s most impactful contribution, his 2020 paper on "Dynamic Parameter Estimation Utilizing Optimized Trajectories" (19 citations), introduced a novel procedure for synthesizing optimal manipulation trajectories. By leveraging parameter aggregates, his method ensures numerically well-conditioned data sets, enabling highly accurate parameter estimation for serial robot manipulators—a foundational advance for precision robotics. More recently, Ulmen has pushed into cutting-edge territory with his 2025 work on Joint Embedding Predictive Architectures (JEPAs), proposing a technique to learn continuous-time state-space models from arbitrary observation data, moving beyond traditional reconstruction-based methods. His 2026 paper, "COM-PACT," addresses the critical need for lightweight neural network controllers in mobile robots and IoT devices, introducing component-aware pruning for latent space models. With a growing citation footprint and a trajectory of innovation, Ulmen is shaping the future of intelligent, efficient autonomous systems.