GRUtopia 2.0 - A general-purpose embodied intelligent simulation platform launched by Shanghai AI Lab

What is GRUtopia 2.0?

GRUtopia 2.0 is a general-purpose embodied intelligence simulation platform released by the Shanghai Artificial Intelligence Laboratory. Building upon GRUtopia 1.0, the platform has undergone a comprehensive upgrade, featuring three core innovations: a general modular framework, automated scene asset generation, and a high-efficiency data acquisition system. Users can define complex tasks with just three lines of code, eliminating the need to switch between multiple platforms. The platform integrates millions of standardized object assets, enabling "one-click generation" of complex scenes and reducing development costs. GRUtopia 2.0 supports various embodied intelligence tasks, such as navigation, manipulation, and motion control. As the core of the "virtual-real integration" technology system, the platform promotes the transition of embodied intelligence from the virtual to the real through high-performance simulation.

Main features of GRUtopia 2.0

• General Modular Simulation FrameworkGRUtopia 2.0 introduces a general modular simulation framework that supports various embodied tasks such as navigation, operation, and motion control. Users can define any task with just three lines of code, eliminating the need to switch between multiple platforms and greatly simplifying the development process.
• Automated generation of scene assetsThe platform integrates millions of standardized object assets and, combined with automated generation and randomization tools, can achieve "one-click generation" of complex scenes.
• High-efficiency data acquisition systemGRUtopia 2.0 offers a variety of easy-to-use teleoperation tools for operational tasks and batch data acquisition tools for navigation tasks. Compared to traditional methods, teleoperation efficiency is improved by 5 times, and data acquisition efficiency for navigation tasks is improved by up to 20 times.
• Large-scale interactive 3D scene datasets (GRScenes)The platform contains 100,000 highly interactive and finely labeled scenes that can be freely combined to create city-scale environments. These scenes cover 89 different categories, filling a gap in the service-oriented environment landscape.
• NPC System (GRResidents)GRUtopia 2.0 introduces an NPC system driven by a Large Language Model (LLM), responsible for social interaction, task generation, and task assignment. It simulates social scenarios, providing a new dimension for embodied AI applications.
• Benchmarking platform (GRBench)The platform introduces GRBench, which supports a variety of robots, especially legged robots as the main intelligent agents, to evaluate their ability to perform medium-difficulty tasks such as object navigation, social navigation, and mobile manipulation.
• Simulation to Reality ParadigmGRUtopia 2.0 demonstrates how simulation can alleviate the scarcity of high-quality data and promote the expansion and application of robotics from the virtual to the real world through a simulation platform.
• Support diverse robotsThe platform supports various types of robots, including humanoid robots and legged robots, and can meet multi-level research needs from low-level control to high-level decision-making.

GRUtopia 2.0 project address

• Github repository:https://github.com/OpenRobotLab/GRUtopia
• arXiv technical paper:https://arxiv.org/pdf/2407.10943

Application scenarios of GRUtopia 2.0

• Robot Training and DevelopmentGRUtopia 2.0 provides a general-purpose modular simulation framework that supports various embodied tasks such as navigation, manipulation, and motion control. Developers can define tasks with just three lines of code, without having to switch between multiple platforms.
• Complex scene constructionThe platform integrates millions of standardized object assets and, combined with automated scene generation and randomization tools, enables "one-click generation" of complex scenes. Scenes cover a variety of environments, including homes, restaurants, offices, and public places, providing robots with diverse training scenarios.
• Social interaction and task generationGRUtopia 2.0 introduces an NPC system driven by a large language model (LLM) that can simulate social interactions, task generation, and task assignment.
• Data Acquisition and OptimizationThe platform provides an efficient data acquisition system that supports a variety of low-threshold remote operation tools for operational tasks and batch data acquisition tools for navigation tasks.