Bart Van Marum

About

Bart Van Marum is a rising leader in legged robotics, whose work sits at the critical intersection of reinforcement learning, computer vision, and bipedal locomotion. His most-cited paper, "Learning Vision-Based Bipedal Locomotion for Challenging Terrain" (2024, 32 citations), tackles a fundamental limitation in the field: the failure of blind, proprioception-only controllers on complex terrain. Van Marum’s key contribution is a framework that integrates visual perception directly into the learning loop, enabling bipedal robots to anticipate and adapt to local obstacles rather than reacting blindly. This work bridges the gap between robust, blind gaits and the environmental awareness needed for real-world deployment. While his citation count is still growing, the impact is already clear—his research provides a practical, learning-based pathway toward truly autonomous bipedal navigation. By addressing the "vision-blind" bottleneck, Van Marum is helping to unlock the next generation of humanoid robots capable of operating in unstructured, human-centric environments, from disaster response to household assistance.

Research Focus

Key Achievements

Top Papers

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

   32 citations · 2024

Key Collaborators

Contact & Links