Cassie sim-to-real paper learns all common bipedal gaits

Key points

• The paper proposes periodic reward composition for standing, walking, hopping, running, and skipping without requiring reference motions.
• The authors demonstrate sim-to-real transfer of learned gaits to Cassie and a generic policy that transitions between all two-beat gaits.
• It is an important Cassie research entry because it treats the robot as a broad bipedal gait-learning platform rather than a single walking demo.