Skip to main content

Bipedal Locomotion and Balance Control

Introduction​

Bipedal locomotion, or walking on two legs, is a hallmark of humanoids and a significant challenge in robotics. Unlike wheeled or quadrupedal robots, bipedal robots are inherently unstable and require sophisticated control strategies to maintain balance and achieve stable, efficient movement. This section explores the principles and techniques behind bipedal locomotion and balance control in humanoid robots.

Principles of Bipedal Locomotion​

Center of Mass (CoM) and Zero Moment Point (ZMP)​

  • CoM: The average position of all the mass in the robot. For stable walking, the projection of the CoM onto the ground should generally remain within the support polygon.
  • ZMP: As discussed in kinematics and dynamics, the ZMP is the point on the ground where the net moment of all forces is zero. For static or quasi-static walking (slow, controlled movements), the ZMP must stay within the support polygon to prevent tipping.

Support Polygon​

  • The area on the ground defined by the points of contact between the robot's feet and the ground. During walking, the support polygon changes as feet lift and make contact.

Walking Patterns and Trajectory Generation​

Static Walking​

  • Concept: The robot's ZMP always remains within the support polygon. This results in slow, deliberate movements, often characterized by distinct phases of single and double support.
  • Method: Each step is planned such that the robot is always stable.

Dynamic Walking​

  • Concept: The ZMP is allowed to move outside the support polygon for brief periods, relying on the robot's inertia to maintain balance. This enables faster, more natural-looking gaits.
  • Method: More complex control algorithms are needed to manage momentum and ensure stability. Often involves predicting the ZMP and adjusting foot placement or torso lean.

Trajectory Generation​

  • Method: Generating smooth, continuous paths for the robot's CoM, ZMP, and joint angles to execute a walking step.
  • Techniques: Often involves optimization techniques to minimize energy consumption or maximize walking speed while maintaining stability.

Balance Control Strategies​

Ankle Strategy​

  • Concept: Primarily used for small disturbances when the robot is standing. The robot applies torques at the ankle joints to shift the CoM and maintain balance.
  • Analogy: Similar to how humans sway slightly to maintain balance.

Hip Strategy​

  • Concept: Used for larger disturbances or when the ankle strategy is insufficient. The robot bends at the hips to shift the CoM.
  • Analogy: Leaning forward or backward from the waist.

Step Strategy​

  • Concept: If the robot loses balance beyond the limits of ankle and hip strategies, it takes a step to create a new support polygon and regain stability.
  • Analogy: Taking a corrective step to avoid falling.

Key Challenges in Bipedal Locomotion​

  • Energy Efficiency: Designing gaits that are energy-efficient for long-duration operation.
  • Rough Terrain: Adapting walking patterns to uneven, slippery, or deformable surfaces.
  • Disturbances: Robustly responding to external pushes or unexpected changes in the environment.
  • Human-like Gaits: Achieving natural-looking and smooth walking patterns that mimic human motion.

Research and Development​

Active research areas include:

  • Reinforcement learning for dynamic walking.
  • Adaptive control for uncertain environments.
  • Human-inspired control architectures.
  • Utilizing whole-body control frameworks.