Simulation of Upward Jump Control for One-Legged Robot Based on QP Optimization

Area of Science:

• Robotics
• Control Systems
• Biomechanics

Background:

• Generating dynamic humanoid motions like jumping is challenging.
• Existing methods struggle to simultaneously satisfy multiple complex constraints.

Purpose of the Study:

• To develop an optimization framework for robust upward jumping motion.
• To integrate trajectory generation and real-time control for jumping.
• To ensure adherence to critical constraints such as ZMP and anti-slippage.

Main Methods:

• Discretized continuous trajectories for launch phase into a nonlinear optimization problem.
• Employed quadratic programming (QP) for real-time control during stance phase.
• Unified control objectives (CoM, angle, angular momentum) and constraints (anti-slippage, ZMP, joint acceleration).

Main Results:

• Achieved a successful upward jump simulation of 16.4 cm height.
• Demonstrated full satisfaction of all imposed control objectives and constraints.
• Validated the effectiveness of the QP-based optimization framework.

Conclusions:

• The proposed QP optimization framework enables successful and constrained upward jumping motions.
• The integrated approach addresses limitations in trajectory generation and real-time control.
• This method provides a robust solution for dynamic robotic locomotion.