Skip to main content

Physics Simulation in Robotics

Objective​

This chapter introduces students to the fundamentals of physics simulation in robotics, focusing on how to accurately model real-world physical phenomena like gravity, collisions, and friction in simulation environments.

Learning Outcomes​

After completing this chapter, students will be able to:

  1. Configure accurate physics properties to match real-world robot behavior in simulation
  2. Understand and implement key physics parameters like gravity, friction, and collision models
  3. Recognize the impact of physics parameters on robot simulation accuracy and performance

Theory​

Physics simulation in robotics is a critical component of creating accurate digital twins that properly mirror real-world robot behavior. The goal is to model the forces and interactions that govern how robots move, interact with objects, and respond to their environment.

Gravity Simulation​

Gravity is the most fundamental force in physics simulation. In Gazebo and similar simulators, gravity is typically represented as a constant vector, usually set to Earth's gravitational acceleration (-9.8 m/s² in the Z direction). However, different environments might require different gravity settings, such as reduced gravity for lunar robots or altered direction for inverted pendulum systems.

Collision Detection and Response​

Collision detection determines when objects in simulation come into contact with each other. Simulators use various algorithms to compute collision points, penetration depths, and response forces. Common approaches include bounding volume hierarchies, spatial hashing, and swept volumes.

Collision response then calculates the resulting forces and torques when objects collide. This involves:

  • Contact point determination
  • Contact normal calculation
  • Friction modeling
  • Restitution (bounciness) handling
  • Impulse-based or force-based response methods

Friction Models​

Friction models determine how surfaces interact when sliding against each other. The Coulomb friction model is commonly used, which includes:

  • Static friction: The force required to initiate motion
  • Dynamic friction: The force required to maintain motion
  • Viscous friction: Velocity-dependent friction

Inertial Properties​

Accurate inertial properties are crucial for realistic physics simulation. These include:

  • Mass: Total amount of matter in the object
  • Center of Mass: Point where mass is concentrated
  • Inertia Tensor: How mass is distributed with respect to rotation

Physics Engines​

Different simulators use different physics engines, each with specific strengths:

  • ODE (Open Dynamics Engine): Good for general rigid body simulation, handles complex joints well
  • Bullet: Excellent collision detection, good for real-time applications
  • Simbody: Designed for biomechanics and complex multi-body systems
  • DART: Combines features of ODE and Simbody

Simulation Accuracy vs. Performance​

There's always a trade-off between simulation accuracy and computational performance. Factors that affect this include:

  • Time step size: Smaller steps increase accuracy but decrease performance
  • Solver iterations: More iterations improve accuracy but increase computation time
  • Constraint handling: More robust constraint handling increases fidelity but reduces speed

Practical Examples​

Example 1: Configuring Physics in Gazebo World File​

Setting up fundamental physics properties in a Gazebo world file:

<world>
  <physics name="ode_physics" type="ode">
    <gravity>0 0 -9.8</gravity>
    <ode>
      <solver>
        <type>quick</type>
        <iters>1000</iters>
        <sor>1.3</sor>
      </solver>
      <constraints>
        <cfm>0.0</cfm>
        <erp>0.2</erp>
        <contact_max_correcting_vel>100.0</contact_max_correcting_vel>
        <contact_surface_layer>0.001</contact_surface_layer>
      </constraints>
    </ode>
  </physics>

  <!-- Robot model with specified physical properties -->
  <model name="robot">
    <pose>0 0 0.5 0 0 0</pose>
    <link name="chassis">
      <collision name="collision">
        <geometry>
          <box><size>1 1 0.5</size></box>
        </geometry>
        <surface>
          <friction>
            <ode>
              <mu>1.0</mu>
              <mu2>1.0</mu2>
              <fdir1>0 0 1</fdir1>
              <slip1>0.0</slip1>
              <slip2>0.0</slip2>
            </ode>
          </friction>
          <bounce>
            <restitution_coefficient>0.0</restitution_coefficient>
            <threshold>100000</threshold>
          </bounce>
          <contact>
            <ode>
              <soft_cfm>0.0</soft_cfm>
              <soft_erp>0.2</soft_erp>
              <kp>1e+13</kp>
              <kd>1.0</kd>
              <max_vel>100.0</max_vel>
              <min_depth>0.0</min_depth>
            </ode>
          </contact>
        </surface>
      </collision>
      <visual name="visual">
        <geometry>
          <box><size>1 1 0.5</size></box>
        </geometry>
      </visual>
      <inertial>
        <mass>10.0</mass>
        <inertia>
          <ixx>1.0</ixx>
          <ixy>0.0</ixy>
          <ixz>0.0</ixz>
          <iyy>1.0</iyy>
          <iyz>0.0</iyz>
          <izz>2.0</izz>
        </inertia>
      </inertial>
    </link>
  </model>
</world>

Example 2: Physics Parameter Tuning for Accuracy​

Fine-tuning physics parameters to match real-world robot behavior:

<!-- For a robot that needs precise manipulation -->
<physics name="precise_physics" type="ode">
  <gravity>0 0 -9.8</gravity>
  <max_step_size>0.001</max_step_size>  <!-- Small time step for precision -->
  <real_time_update_rate>1000</real_time_update_rate>  <!-- High update rate -->
  <ode>
    <solver>
      <iters>2000</iters>  <!-- Increase iterations for better convergence -->
      <sor>1.2</sor>
    </solver>
    <constraints>
      <contact_surface_layer>0.0001</contact_surface_layer>  <!-- Minimal contact layer -->
      <contact_max_correcting_vel>10.0</contact_max_correcting_vel>  <!-- Limit correction velocity -->
    </constraints>
  </ode>
</physics>

Hands-on Lab​

Prerequisites​

  • Gazebo installed and running
  • Understanding of basic SDF world files
  • Knowledge of robot models from previous chapters

Step 1: Create a Physics Test Environment​

Create a world file with basic physics parameters and add objects with different properties (sphere, box, cylinder) to observe their behavior.

Step 2: Modify Gravity Parameters​

Adjust the gravity vector in the world file to see how it affects object behavior. Try setting gravity to (0, 0, -5) and observe the difference from standard Earth gravity (0, 0, -9.8).

Step 3: Adjust Friction Coefficients​

Create objects with different friction coefficients (low μ = 0.1, high μ = 1.0) and observe how they slide against surfaces.

Step 4: Tune Collision Parameters​

Experiment with different ERP (Error Reduction Parameter) and CFM (Constraint Force Mixing) values to see how they affect:

  • Stability of stacked objects
  • Penetration between objects
  • Overall simulation fidelity

Step 5: Inertial Properties Experiment​

Create two identical-looking objects but with different inertial properties and observe how they react differently to forces.

Exercises​

  1. Design a physics configuration that would best simulate a robot on a low-friction ice surface. What specific parameters would you adjust?

  2. Explain why using extremely small time steps in physics simulation might not always yield better results.

  3. Compare the physics parameters needed for simulating a fast-moving wheeled robot versus a slow-moving manipulator arm.

  4. Describe what happens when the restitution coefficient is set to 0, 0.5, and 1.0, and give examples of materials that might have these properties.

  5. Research and explain the difference between static and dynamic friction, and describe a scenario where this distinction matters in robotics.

Summary​

This chapter covered the fundamentals of physics simulation in robotics, emphasizing the importance of accurate parameter configuration. Students learned about key physics concepts including gravity, collision detection, friction models, and inertial properties. The chapter also discussed the balance between simulation accuracy and computational performance, and provided practical examples of physics configuration in Gazebo.

Further Reading​