Overview

GRID provides users the access to Isaac Sim for large-scale data generation, evaluation, and training for robot learning. It supports a wide range of robots including manipulators, quadrupeds, and humanoids—across diverse environments such as kitchens, offices, industrial areas, and outdoor spaces.

​

Key Features

• Isaac Sim 4.1 Ready: GRID builds on top of NVIDIA Isaac Sim, enabling you to leverage the latest features in Isaac Sim. We’re in the process of migrating to IsaacSim 4.5 and it will be available soon to all GRID users.

• Parallelized RL Training: Support for training thousands of parallel agents in simulation for a number of diverse tasks.

• Large-Scale Data Generation: Generate high-quality simulation data at scale using VR teleoperation, motion planning, and reinforcement learning.

• Advanced Locomotion and Manipulation Capabilities: Access pre-trained policies for various robots, enabling complex movements and interactions across different environments.