CRL Robot Parkour

This summer, Taerim Yoon from Korea University joined the Computational Robotics Lab through the ETH Robotics Student Fellowship program. Over the last two months, Taerim developed a "Parkour Controller" for quadruped robots using reinforcement learning, enabling them to exhibit extremely agile behaviors over challenging obstacles. We are excited to see how far we can push this to achieve even more agile and robust legged locomotion behaviors. Please stay tuned! Acknowledgment: - This work is a collaboration between Taerim Yoon from the Robot Intelligence Lab, Korea University and Dongho Kang from the Computational Robotics Lab, ETH Zurich. - Flavio De Vincenti and Hélène Stefanelli contributed to the development of the control pipeline. - Video courtesy of Taerim Yoon, Jin Cheng, and Dongho Kang ETH Robotics Student Fellowship: https://center-for-robotics.ethz.ch/education/robotics-student-fellowship.html Computational Robotics Lab: https://crl.ethz.ch/ Robot Intelligence Lab: https://sites.google.com/view/sungjoon-choi/home Taerim Yoon: - https://www.linkedin.com/in/taerim-yoon-631a9321a/ Dongho Kang: - https://donghok.me/ - https://twitter.com/eastskykang Jin Cheng: - https://jin-cheng.me/ - https://twitter.com/catachiii