🤖

Robotics & Simulation Lab

GCP Compute Engine with NVIDIA GPU + NVIDIA Isaac Sim

Overview

The Robotics & Simulation Lab brings robotics into the cloud. Instead of needing a physical robot on day one, you work in high-fidelity simulators — NVIDIA Isaac Sim for photorealistic environments and Gazebo for lightweight ROS 2 development. The simulated robots have the same sensors (cameras, LiDAR, IMU), the same motor controllers, and the same ROS 2 interfaces as their physical counterparts. Train a perception model, develop a navigation stack, or teach a robot arm to grasp objects with reinforcement learning — all in simulation before touching real hardware.

What You'll Do in This Lab

  • Set up ROS 2 workspaces and launch robots in Gazebo simulation
  • Build perception pipelines — process LiDAR point clouds, run object detection on camera feeds
  • Implement SLAM (Simultaneous Localization and Mapping) in simulated environments
  • Train reinforcement learning agents for robot control with Stable Baselines3
  • Build autonomous navigation stacks using Nav2
  • Generate synthetic training data with domain randomization in Isaac Sim

Lab Workflow

1

Launch

Start the lab. A GPU instance with Isaac Sim, Gazebo, and ROS 2 Humble launches. A VNC/web session gives you the graphical interface.

2

Load Scene

Open a simulation scenario — warehouse with mobile robots, tabletop with a robot arm, or outdoor environment with a drone.

3

Develop

Write ROS 2 nodes in the integrated terminal. Run perception, planning, or control algorithms while the simulator provides sensor data.

4

Train

For RL exercises, launch training in the simulator. The agent interacts with the environment thousands of times to learn optimal behavior.

5

Evaluate

Run evaluation episodes. Measure success rates, collision counts, path efficiency, and task completion time.

6

Export

Save trained policies and perception models. Transfer to the Edge AI Lab for deployment on physical Jetson hardware.

Hardware & Environment

Machine Type n1-standard-16 (16 vCPU, 60 GB RAM)
GPU 1x NVIDIA T4 (required for Isaac Sim rendering)
Simulators NVIDIA Isaac Sim 2023.1+, Gazebo Fortress
ROS ROS 2 Humble with full desktop install
Interface Web-based VNC for GUI access (noVNC)
Session Length 2-4 hour sessions

Pre-installed Tools

NVIDIA Isaac Sim Gazebo + ROS 2 Stable Baselines3 (RL) Open3D (point clouds) Nav2 (navigation)

Frequently asked questions about this lab

What is the Robotics & Simulation Lab? +
High-performance simulation environment for robotics AI. Students work with physics simulators, train RL agents, and develop perception and navigation systems in realistic virtual environments.
Which courses use this lab? +
This lab is included in: AI for Robotics & Edge Computing.
What hardware does this lab run on? +
GCP Compute Engine with NVIDIA GPU + NVIDIA Isaac Sim. Machine Type: n1-standard-16 (16 vCPU, 60 GB RAM); GPU: 1x NVIDIA T4 (required for Isaac Sim rendering); Simulators: NVIDIA Isaac Sim 2023.1+, Gazebo Fortress; ROS: ROS 2 Humble with full desktop install.
What software comes pre-installed? +
Comes pre-loaded with NVIDIA Isaac Sim, Gazebo + ROS 2, Stable Baselines3 (RL), Open3D (point clouds), Nav2 (navigation). No local installs or dependency setup required — open your browser and start working.
Can I bring my own datasets and code into this lab? +
Yes. Datasets can be uploaded directly or synced from Google Cloud Storage. Notebooks and source files have built-in Git integration so you can push work to your own GitHub or GitLab repos.
Do I need to enroll in a course to use this lab? +
Yes. Lab environments are provisioned per-student as part of an AI Labs course enrollment. Browse the courses linked above to find programs that include this lab.

Related labs

Other AI Labs environments students typically use alongside this one.

Ready to Try This Lab?

Enroll in a course that uses this lab, or visit our Houston center for a hands-on demo.