Whole-Body Dynamics-Based Aerial Fall Trajectory Optimization and Landing Control for Humanoid Robot

Area of Science:

• Robotics
• Control Systems
• Mechanical Engineering

Background:

• Falls are common for humanoid robots in complex human environments.
• Existing research primarily addresses ground falls, neglecting aerial falls.
• Aerial falls pose significant risks to robot integrity and safety.

Purpose of the Study:

• To develop and validate methods for safe aerial falls in humanoid robots.
• To optimize falling trajectories and mitigate impact forces.
• To enhance robot resilience and safety in dynamic environments.

Main Methods:

• Utilized an extended state variable formulation for trajectory optimization.
• Employed proportional-differentiation (PD) control during the aerial phase.
• Optimized contact forces using a centroidal dynamics model and a dual spring-damper model for landing.

Main Results:

• Simulations demonstrated safe falls from 1.5 meters at a 45° pitch angle.
• Experimental validation on a physical robot confirmed the effectiveness of the methods.
• Achieved excellent shock absorption for impact forces during aerial falls.

Conclusions:

• The proposed trajectory optimization and motion control effectively protect humanoid robots during aerial falls.
• These methods significantly reduce impact forces, enhancing robot durability.
• Enables safer operation of humanoid robots in unpredictable human environments.