Papers
3
Total Citations
155
H-Index
2
About
Marlon Gwira is at the forefront of applying deep reinforcement learning to create agile, intelligent behaviors for humanoid robots, with a particular focus on the demanding domain of robot soccer. His research demonstrates that complex, dynamic skills—such as playing a simplified one-versus-one game—can be synthesized for low-cost, miniature humanoid platforms using deep RL. Gwira’s most cited work, “Learning agile soccer skills for a bipedal robot with deep reinforcement learning” (2024), has already garnered 147 citations, highlighting its significant impact on the field. He has further pushed boundaries by training robots to play soccer using only egocentric RGB vision and onboard computation, tackling real-world challenges like active perception and long-horizon planning. By proving that sophisticated, safe movement strategies can be composed from learned skills, Gwira’s work bridges the gap between simulation and physical deployment, offering a powerful framework for developing versatile, autonomous robots capable of operating in dynamic, unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3