SIMULTANEOUS PLANNING AND DYNAMIC STABILITY SIMULATION VERIFICATION OF HUMANOID ROBOT DANCE MOVEMENTS

Abstract

This paper addresses the problem of multi-joint trajectory planning and cooperative control for humanoid robots in dance performances. Taking the dance task of the Unitree G1 robot developed by Hangzhou Unitree Robotics at the opening ceremony of a technology exhibition as the background, a hierarchical motion-planning method is proposed that decomposes the overall movement into three sub-tasks: torso rotation, coordinated arm motion, and leg balance adjustment. Torso rotation is planned with a fifth-order polynomial for smooth trajectories; the two arms perform counter-rotating circular motions and their joint angles are computed via inverse kinematics; the legs provide real-time balance compensation based on centroidal dynamics and the Zero-Moment Point (ZMP) stability criterion. Simulation analysis yields coordinated motion trajectories for all 12 degrees of freedom of the torso, arms and legs, achieving fluent dance motions while guaranteeing overall stability and safety during dynamic movement. The study offers a viable theoretical and methodological reference for motion planning and control of humanoid robots in complex scenarios such as artistic performances.