Bart Jaap van Marum

About

Bart Jaap van Marum is a leading researcher at the intersection of reinforcement learning and legged robotics, with a primary focus on vision-based bipedal locomotion for challenging terrains. His most cited work, "Learning Vision-Based Bipedal Locomotion for Challenging Terrain" (2023, 2 citations), addresses a critical gap in robotic control: while blind proprioceptive controllers can handle moderate terrains, they fail in environments requiring anticipation and adaptation to local features. Van Marum’s major contribution lies in integrating visual perception with RL-based locomotion policies, enabling bipedal robots to dynamically adjust their gaits based on terrain ahead. This work has significant implications for search-and-rescue, planetary exploration, and disaster response, where robots must navigate unpredictable landscapes. Though early in his career, his research has already garnered attention for its novel approach to bridging perception and control. Van Marum’s achievements include developing robust training frameworks that allow robots to generalize from simulation to real-world deployment, a key challenge in robotics. His ongoing work continues to push the boundaries of autonomous locomotion, promising safer and more adaptable robots for complex environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning Vision-Based Bipedal Locomotion for Challenging Terrain

   2 citations · 2023

Key Collaborators

Contact & Links