CyberDog 2 Dribbling with Deep Reinforcement Learning
Jun 1, 2024
·
1 min read
Trained quadrupedal and bipedal dribbling policies for CyberDog 2 using reinforcement learning.
Implemented a vision-based goalkeeper module and integrated a complete 2v2 match environment. The system achieved 2nd place out of 16 teams in a university robotics competition.
Authors
B.S. Candidate in Computer Science
I am a CS undergraduate at Tsinghua IIIS working on embodied AI,
Vision-Language-Action models, and robotic control. My recent work studies
how agents connect language, perception, memory, and continuous action in
simulated and physical environments.
Email: zhang-zz22@mails.tsinghua.edu.cn
WeChat: zzz1607432325