Online Centroidal Angular Momentum Reference Generation and Motion Optimization for Humanoid Push Recovery

This letter presents a new push recovery algorithm for humanoid robots in balancing scenarios by exploiting the system's rotational dynamics. The proposed framework actively generates centroidal angular momentum (CAM) references based on force magnitude and direction of to counteract disturbance maintain its balance. Since robot can only store limited amount momentum, CAM reference is generated three consecutive phases: 1) generation phase push; 2) reduction bring halt; 3) posture converge pose again. A subsequent whole-body motion optimizer, formulated as constrained quadratic optimization problem, kinematically feasible trajectories reference. validated through experiments with TORO.