Heechan Shin gave a presentation concerning his work at RSS 2023 workshop (Experiment-oriented Locomotion and Manipulation Research), which was held in Daegu, South Korea. The paper proposes a quadruped fault-tolerant locomotion using curriculum learning.
