Gyro-Sensor-Based Vibration Control for Dynamic Humanoid-Robot Walking on Inclined Surfaces

Area of Science:

• Robotics
• Control Systems
• Humanoid Robot Dynamics

Background:

• Stable walking of humanoid robots on inclined surfaces remains a significant challenge in robotics.
• Existing motor-control systems often struggle with dynamic instabilities on uneven terrain.
• The KONDO KHR-3HV lightweight humanoid robot is used as a platform for investigation.

Purpose of the Study:

• To investigate an efficient motor-control system for stable walking of the KONDO KHR-3HV on inclined surfaces.
• To analyze the influence of angular-pitch velocity on robot gait stability.
• To propose a novel frequency-domain analysis method for understanding humanoid robot walking instabilities.

Main Methods:

• Utilized a gyro sensor to detect the angular-pitch velocity of the robot's torso.
• Performed gait analysis on various downslopes with and without motor-feedback control.
• Applied frequency-domain analysis to the angular-pitch velocity data.

Main Results:

• Nonlinear motor torque, induced by slope forces, generates harmonics of the fundamental walking frequency (1.73 Hz), causing instability.
• Feedback-gain parameters (K_A and K_H) influence harmonic amplitudes, leading to vibrations at higher inclinations.
• Increased surface friction allows reduced feedback gain, mitigating harmonic contributions and enhancing stability.

Conclusions:

• Harmonics in angular-pitch velocity are identified as the primary cause of unstable humanoid robot walking on slopes.
• Motor-control system design must consider the interplay between feedback gain, surface friction, and harmonic generation.
• To ensure stability, the damped natural frequency of the motor-control system should be maintained below the fundamental walking frequency.