Chapter 3.2: Locomotion and Manipulation Control

Learning Objectives

By the end of this chapter, you will be able to:

• Analyze the principles of bipedal locomotion and dynamic balancing in humanoids
• Evaluate different manipulation strategies and grasping techniques for humanoid robots
• Assess whole-body control strategies for coordinating locomotion and manipulation
• Design control approaches that balance stability and dexterity in humanoid systems

3.2.1 Dynamic Balancing and Bipedal Locomotion

Bipedal locomotion is a hallmark of humanoids, offering the ability to navigate human-centric environments designed for two-legged beings. However, achieving stable and robust bipedal walking is one of the most challenging problems in robotics due to the inherent instability of an inverted pendulum system.

• Zero Moment Point (ZMP):

  • Concept: The ZMP is a fundamental concept in bipedal locomotion. It is the point on the ground about which the sum of all moments due to gravitational and inertial forces is zero. For stable walking, the ZMP must always remain within the robot's support polygon (the area defined by the contact points of the feet with the ground).
  • Control: ZMP-based controllers calculate the desired ZMP trajectory and generate joint torques or foot placements to keep the actual ZMP close to the desired trajectory, ensuring stability.
  • Limitations: While effective for stable walking, traditional ZMP control can lead to stiff and less dynamic gaits, making it difficult to handle unexpected disturbances or highly uneven terrain.

• Whole-Body Control (WBC):

  • Concept: WBC approaches consider the entire robot's kinematic and dynamic model simultaneously. It optimizes the motion of all joints to achieve multiple objectives (e.g., maintain balance, reach a target, avoid obstacles) while respecting various constraints (e.g., joint limits, contact forces, friction cones).
  • Hierarchical/Prioritized Control: Often, WBC formulates tasks in a hierarchy, where higher-priority tasks (like balance) are guaranteed to be satisfied before lower-priority tasks (like manipulation).
  • Advantages: Enables more dynamic, human-like motions, better disturbance rejection, and simultaneous execution of locomotion and manipulation tasks.
  • Examples: Used in advanced humanoids like HRP series robots and sophisticated research platforms to achieve highly coordinated and adaptive movements.

• Compliant Actuation and Series Elastic Actuators (SEAs):

  • Concept: Integrating elastic elements (springs) into the robot's actuators to provide inherent compliance.
  • Benefits for Locomotion: SEAs allow for more natural, compliant interactions with the ground, absorbing shocks from impacts, and storing/releasing energy for efficient, dynamic movements (e.g., jumping, running). They also make the robot safer for interaction by limiting impact forces.
  • Control: Controlling SEAs requires specialized techniques that account for the spring dynamics.

• Gait Generation:

  • Trajectory Optimization: Algorithms that optimize joint trajectories and footstep placements to generate stable and energy-efficient walking gaits.
  • Model Predictive Control (MPC): Uses a model of the robot and its environment to predict future states and optimize control inputs over a finite horizon, recalculating at each time step. This allows humanoids to react proactively to disturbances and generate adaptive gaits.

tip

The key to successful bipedal locomotion lies in the robot's ability to continuously adjust its center of mass and maintain its balance while moving. This requires sophisticated sensors and real-time control algorithms that can respond quickly to changes in the environment or unexpected disturbances.

3.2.2 Dexterous Manipulation and Grasping

Humanoid hands and arms are designed to replicate human-like dexterity, enabling complex object interaction, tool use, and manipulation in various environments.

• Kinematic Redundancy: Humanoid arms typically have many degrees of freedom (e.g., 7 DoF per arm), allowing for highly flexible and obstacle-avoiding movements. This redundancy is exploited by inverse kinematics algorithms to find optimal joint configurations for desired end-effector poses.
• Multi-Fingered Hands:
  • Design: Mimic the human hand structure, often with multiple fingers, each with several joints. This design enables a wide range of grasps (e.g., power grasp, precision grasp) and dexterous in-hand manipulation.
  • Control: Controlling multi-fingered hands is highly complex, involving coordination of numerous joints to achieve stable grasps, apply appropriate forces, and perform fine manipulation. Algorithms focus on contact force distribution, grasp stability, and tactile feedback integration.
• Grasping Strategies:
  • Form Closure: Achieved when the object is completely constrained by the grippers, preventing any motion.
  • Force Closure: Achieved when the contact forces can resist any external disturbance on the object, even if it's not fully constrained kinematically.
  • Grasp Planning: Involves identifying stable contact points and joint configurations to securely hold an object, considering its geometry, weight, and material properties. This can be pre-programmed or learned through machine learning.
• Perception for Manipulation:
  • Visual Servoing: Using camera feedback to guide the robot's end-effector to a target or track an object during manipulation.
  • Tactile Feedback: Integrating tactile sensors in the grippers to detect contact, pressure, and slippage, allowing for adaptive grasping and delicate manipulation.

3.2.3 Whole-Body Control Strategies

Whole-Body Control (WBC) is a critical paradigm for humanoids, integrating locomotion, balance, and manipulation into a unified control framework. It allows the robot to leverage its entire body to achieve complex tasks more robustly and effectively.

• Integrated Control: WBC addresses the challenge of coordinating the many degrees of freedom of a humanoid robot by formulating a set of tasks (e.g., balance, end-effector position, gaze direction) that are achieved simultaneously or hierarchically.
• Task Prioritization: Tasks are often prioritized. For example, maintaining balance might be a higher priority than precisely reaching a target. If conflicts arise, higher-priority tasks are satisfied first, and remaining control authority is used for lower-priority tasks.
• Dynamic Consistency: WBC ensures that the generated motions are dynamically consistent, meaning they respect the robot's physical properties (mass, inertia) and the laws of physics.
• Contact Management: Crucial for tasks involving multiple contacts with the environment (e.g., walking, pushing against a wall). WBC determines optimal contact forces and ensures they remain within friction limits.
• Applications:
  • Locomotion and Manipulation Simultaneously: A humanoid walking while carrying an object, or opening a door while maintaining balance.
  • Disturbance Rejection: Using arm swings or torso movements to regain balance after an external push.
  • Complex Tool Use: Coordinating hand, arm, and body movements to effectively wield tools.
  • Human-Robot Collaboration: Adjusting posture and force during physical interaction with a human.

WBC represents a significant step towards enabling humanoids to perform truly human-like, versatile tasks in unstructured environments by allowing them to exploit their full physical capabilities.

Try it yourself

Research and compare the locomotion strategies of two different humanoid robots (e.g., ASIMO and Atlas). Analyze how their different approaches to balance control (ZMP vs. MPC) and actuator types (rigid vs. compliant) influence their gait patterns and overall performance in various terrains.