Tutorials — Isaac Lab Documentation (original) (raw)

Contents

• Setting up a Simple Simulation
• Interacting with Assets
• Creating a Scene
• Designing an Environment
• Integrating Sensors
• Using motion generators

Tutorials#

Welcome to the Isaac Lab tutorials! These tutorials provide a step-by-step guide to help you understand and use various features of the framework. All the tutorials are written as Python scripts. You can find the source code for each tutorial in the scripts/tutorials directory of the Isaac Lab repository.

Note

We would love to extend the tutorials to cover more topics and use cases, so please let us know if you have any suggestions.

We recommend that you go through the tutorials in the order they are listed here.

Setting up a Simple Simulation#

These tutorials show you how to launch the simulation with different settings and spawn objects in the simulated scene. They cover the following APIs: AppLauncher,SimulationContext, and spawners.

• Creating an empty scene
• Spawning prims into the scene
• Deep-dive into AppLauncher

Interacting with Assets#

Having spawned objects in the scene, these tutorials show you how to create physics handles for these objects and interact with them. These revolve around the AssetBaseclass and its derivatives such as RigidObject,Articulation and DeformableObject.

• Adding a New Robot to Isaac Lab
• Interacting with a rigid object
• Interacting with an articulation
• Interacting with a deformable object

Creating a Scene#

With the basic concepts of the framework covered, the tutorials move to a more intuitive scene interface that uses the InteractiveScene class. This class provides a higher level abstraction for creating scenes easily.

• Using the Interactive Scene

Designing an Environment#

The following tutorials introduce the concept of manager-based environments: ManagerBasedEnvand its derivative ManagerBasedRLEnv, as well as the direct workflow base classDirectRLEnv. These environments bring-in together different aspects of the framework to create a simulation environment for agent interaction.

• Creating a Manager-Based Base Environment
• Creating a Manager-Based RL Environment
• Creating a Direct Workflow RL Environment
• Registering an Environment
• Training with an RL Agent
• Modifying an existing Direct RL Environment
• Policy Inference in USD Environment

Integrating Sensors#

The following tutorial shows you how to integrate sensors into the simulation environment. The tutorials introduce the SensorBase class and its derivatives such as Camera and RayCaster.

• Adding sensors on a robot

Using motion generators#

While the robots in the simulation environment can be controlled at the joint-level, the following tutorials show you how to use motion generators to control the robots at the task-level.

• Using a task-space controller
• Using an operational space controller