Papers
3
Total Citations
155
H-Index
2
About
Kushal Patel is a leading researcher in robotics and artificial intelligence, specializing in deep reinforcement learning (deep RL) for agile, full-body control of humanoid robots. His most impactful work, "Learning agile soccer skills for a bipedal robot with deep reinforcement learning," has garnered 147 citations, demonstrating its significant influence on the field. In this study, Patel and his team showed that deep RL could synthesize sophisticated, safe movement skills for a low-cost, miniature humanoid robot, enabling it to play a simplified one-versus-one soccer game. This breakthrough proved that complex, dynamic behaviors could be learned and composed from scratch, pushing the boundaries of what is possible with small-scale robotics. Patel further advanced the field by extending this approach to egocentric vision in "Learning Robot Soccer from Egocentric Vision with Deep Reinforcement Learning," where he trained policies using only onboard computation and RGB cameras. This work tackles real-world challenges like active perception and long-horizon planning, marking a major step toward fully autonomous, agile robots. His contributions are foundational for developing robots capable of navigating and interacting in unstructured human environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3