Jan Achterhold
Papers
2
Total Citations
5
H-Index
2
About
Jan Achterhold is an emerging researcher at the intersection of robotics, machine learning, and physical modeling. His work focuses on two compelling domains: autonomous robot navigation under variable conditions, and physics-informed learning systems for dynamic trajectory prediction. In his work on context-conditional navigation, Achterhold addresses a critical challenge in real-world robotics — the fact that terrain properties like friction coefficients and robot dynamics can shift unpredictably due to environmental changes or varying payloads. By developing a learning-based model that remains aware of both terrain and robot characteristics, he advances the field of adaptive autonomous navigation, making robotic systems more robust and generalizable. His table tennis trajectory prediction research showcases his strength in combining physical priors with data-driven learning — a so-called gray-box approach. By integrating an extended Kalman filter with neural spin inference into a physically grounded model, he demonstrates how hybrid methods can outperform purely black-box alternatives in structured physical domains. Though early in citation accumulation — with 3 and 2 citations respectively for his 2023 publications — Achterhold's methodological sophistication and focus on bridging theory with real-world applicability position him as a promising contributor to the robotics and embodied AI research communities.
Research Focus
Key Achievements
Top Papers
- 1
- 2