Robust Fall Recovery for Armless Bipedal-Wheeled Robots Via Force-Guided Learning

The article introduces the Force-guided Teacher-student framework with Stage-wise Rewards (FTSR) for enabling fall recovery in armless bipedal-wheeled robots. This approach employs constrained reinforcement learning to optimize recovery strategies by simulating external forces that correlate with the robot's height, allowing for effective posture stabilization and locomotion without arm support. The successful deployment on a physical robot demonstrates robust recovery capabilities in diverse conditions, highlighting its potential for enhancing the adaptability and resilience of legged robotic systems.