Chapter 2: The Robotic Nervous System (ROS 2)

Module 1: Middleware for Robot Control​

Weeks 3-5: ROS 2 Fundamentals​

Learning Objectives​

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

  • Understand ROS 2 architecture and core concepts
  • Create and manage nodes, topics, services, and actions
  • Build ROS 2 packages with Python using rclpy
  • Bridge Python AI agents to ROS controllers
  • Work with URDF (Unified Robot Description Format) for humanoid robots
  • Configure launch files and manage parameters

2.1 What is ROS 2?​

ROS 2 (Robot Operating System 2) is the industry-standard middleware for robotics development. Think of it as the "nervous system" that connects sensors, actuators, and AI algorithms in a distributed robot architecture.

Why ROS 2?​

Unlike ROS 1, ROS 2 is designed for:

  • Real-time performance: Deterministic communication for safety-critical systems
  • Multi-robot systems: Native support for swarms and collaborative robots
  • Security: Built-in encryption and authentication
  • Cross-platform: Linux, Windows, macOS support
  • Industrial deployment: Used by companies like Boston Dynamics, NASA, and Tesla

ROS 2 Distributions​

  • Humble Hawksbill (LTS): Ubuntu 22.04, recommended for this course
  • Iron Irwini: Latest stable release
  • Rolling Ridley: Bleeding-edge development

Core Concepts​

Mechanical Fundamentals​

Degrees of Freedom (DOF)​

A degree of freedom is an independent direction of motion. Human-like robots typically have:

  • Arms: 7 DOF per arm (shoulder: 3, elbow: 1, wrist: 3)
  • Legs: 6 DOF per leg (hip: 3, knee: 1, ankle: 2)
  • Torso: 3 DOF (pitch, roll, yaw)
  • Head: 2-3 DOF (pan, tilt, optional roll)

Total: 25-30 DOF for a full humanoid

Kinematic Chains​

  • Serial chain: Joints connected in sequence (like an arm)
  • Parallel chain: Multiple actuators control a single joint (increased strength)
  • Closed-loop chain: Forms a loop (used in legs for stability)

Actuator Technologies​

1. Electric Motors​

Advantages:

  • Precise position control
  • High repeatability
  • Easy to integrate with digital controllers

Types:

  • DC brushless motors: High efficiency, long lifespan
  • Servo motors: Built-in position feedback
  • Stepper motors: Precise incremental motion

2. Hydraulic Actuators​

Advantages:

  • High power-to-weight ratio
  • Suitable for heavy-duty applications

Disadvantages:

  • Requires hydraulic pump and fluid management
  • Maintenance-intensive
  • Risk of leaks

3. Pneumatic Actuators​

Advantages:

  • Compliant (safe for human interaction)
  • Lightweight

Disadvantages:

  • Lower precision than electric motors
  • Requires compressed air source

Sensors for Perception and Control​

Vision Sensors​

  • RGB cameras: Color imaging for object recognition
  • Depth cameras (e.g., RealSense): 3D environment mapping
  • Stereo cameras: Depth perception through parallax

Proprioceptive Sensors​

  • Encoders: Measure joint angles (position feedback)
  • IMU (Inertial Measurement Unit): Detects orientation and acceleration
  • Force/Torque sensors: Measure interaction forces at joints

Tactile Sensors​

  • Pressure sensors: Detect contact and grip force
  • Skin sensors: Distributed touch sensing on robot surface

Kinematics and Dynamics​

Forward Kinematics​

Given joint angles, calculate the end-effector position:

Position = f(θ₁, θ₂, ..., θₙ)

Example: With 3 joint angles (shoulder, elbow, wrist), determine where the hand is in 3D space.

Inverse Kinematics​

Given desired end-effector position, calculate required joint angles:

(θ₁, θ₂, ..., θₙ) = f⁻¹(desired position)

Challenge: Often has multiple solutions or no solution (workspace limits).

Dynamics​

Dynamics govern how forces and torques affect motion. The equation of motion:

τ = M(q)q̈ + C(q,q̇)q̇ + G(q)

Where:

  • τ: Joint torques (control input)
  • M(q): Inertia matrix
  • C(q,q̇): Coriolis and centrifugal forces
  • G(q): Gravity forces

Practical Application​

Walking Control​

Humanoid walking requires:

  1. Gait planning: Define foot trajectories and center of mass motion
  2. Balance control: Keep center of pressure within support polygon
  3. Compliance: Absorb impacts and adapt to uneven terrain

Zero Moment Point (ZMP) is a key stability criterion: the point where the net moment from gravity and inertia is zero.

Example: Simple Balance Controller​

# Pseudocode for balance control
def balance_controller(imu_data, target_orientation):
    error = target_orientation - imu_data.orientation
    torque_correction = PID_control(error)
    apply_torque_to_ankle_joints(torque_correction)

This simple controller adjusts ankle torques to maintain upright posture.

Summary​

Humanoid robots combine mechanical design, actuation, sensing, and control to achieve human-like motion. The field draws on mechanical engineering, control theory, and AI to create machines that can navigate and interact with human environments.

Key Takeaways:

  • Humanoids have 25-30 degrees of freedom to mimic human motion
  • Actuator choice (electric, hydraulic, pneumatic) involves trade-offs in power, precision, and safety
  • Sensors provide feedback for perception (vision) and control (encoders, IMUs)
  • Kinematics and dynamics are essential for planning and executing motion

Further Reading​

  • Books:

    • Humanoid Robotics: A Reference by Ambarish Goswami and Prahlad Vadakkepat
    • Introduction to Robotics: Mechanics and Control by John J. Craig

  • Research Papers:

    • "Bipedal Walking Control Based on Capture Point Dynamics" (IROS 2011)
    • "Atlas: A Hydraulic Humanoid Robot" by Boston Dynamics

  • Online Resources: