Online Running-Gait Generation for Bipedal Robots with Smooth State Switching and Accurate Speed Tracking

Area of Science:

• Robotics
• Control Engineering
• Biomechanics

Background:

• Bipedal robots require stable and reactive running for practical applications.
• Previous research has struggled to integrate online state switching and speed tracking for running gaits.
• Achieving smooth transitions and accurate speed control remains a significant challenge in bipedal locomotion.

Purpose of the Study:

• To develop an online running-gait generator for bipedal robots.
• To enable smooth state switching (starting/stopping) and accurate online speed tracking.
• To enhance the stability and reactivity of bipedal robots during running.

Main Methods:

• Utilized a simplified variable-height inverted-pendulum (VHIP) model for computational efficiency and to account for fluctuating robot height.
• Proposed a segmented zero moment point (ZMP) trajectory optimization for feasible and smooth center-of-mass (CoM) trajectories during state transitions.
• Developed an iterative algorithm to compute target footholds for accurate online desired speed tracking within two steps.

Main Results:

• Demonstrated smooth and stable starting and stopping of running gaits through ZMP trajectory optimization.
• Achieved accurate online tracking of desired running speeds using the iterative foothold computation method.
• Validated the effectiveness of the proposed methods through numerical experiments and simulations on the BHR7P robot.

Conclusions:

• The proposed online running-gait generator successfully integrates smooth state switching and accurate speed tracking for bipedal robots.
• The VHIP model, ZMP optimization, and iterative foothold computation provide a robust framework for dynamic bipedal running.
• The findings pave the way for more agile and responsive bipedal robots capable of variable-speed locomotion.