robosuite: A Modular Simulation Framework and Benchmark for Robot Learning

By Yuke Zhu, Josiah Wong, Ajay Mandlekar, Roberto Mart'in-Mart'in. 2020

702 citations recorded by the source corpus.

robosuite is a simulation framework for robot learning powered by the MuJoCo physics engine. It offers a modular design for creating robotic tasks as well as a suite of benchmark environments for reproducible research. This paper discusses the key system modules and the benchmark environments of our new release robosuite v1.0.

Subfields: Artificial Intelligence, Control and Systems Engineering, Mechanical Engineering, Evaluation & Benchmarks, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

In Mechanical Engineering

In Evaluation & Benchmarks

In Robotics & Control

Citations in the corpus

A Survey of Embodied AI: From Simulators to Research Tasks (2021, background): “With the advancement in computer graphics and state-of-theart physics engines [110, 111], current embodied AI simulators have all reached a level that separates them from just a conventional game-based simulation for reinforcement learning.”
Generalization Through Hand-Eye Coordination: An Action Space for Learning Spatially-Invariant Visuomotor Control (2021, methodology): “All tasks are designed using MuJoCo [36] and the robo-suite framework [46].”
Learning Multi-Arm Manipulation Through Collaborative Teleoperation (2020, methodology): “Tasks: All tasks were designed using MuJoCo [39] and the robosuite framework [42] (see Fig.”
Human-in-the-Loop Imitation Learning using Remote Teleoperation (2020, methodology): “All tasks were designed using MuJoCo [38] and the robosuite framework [43] (see Fig.”
Can Wikipedia Help Offline Reinforcement Learning? (2022, background): “…et al., 2020), due to the difﬁculty in parameterizing different scenes and tasks through a single network (Wang et al., 2018b; Jiang et al., 2019; Zeng et al., 2020) as well as the lack of large off-the-shelf datasets for pre-training (Cobbe et al., 2020; Zhu et al., 2020; Yu et al., 2020).”
iGibson 2.0: Object-Centric Simulation for Robot Learning of Everyday Household Tasks (2021, neutral): “Recent years, we have seen the emergence of many simulation environments for robotics and 22 embodied AI research [1, 2, 3, 4, 5, 6, 7, 8, 9, 10].”
VIMA: General Robot Manipulation with Multimodal Prompts (2022, methodology): “, 2020) and Robosuite (Zhu et al., 2020; Fan et al., 2021) design various tabletop manipulation tasks with realistic robot arms.”
Modular Lifelong Reinforcement Learning via Neural Composition (2022, neutral): “All dynamics are simulated in robosuite (Zhu et al., 2020).”
Evaluating Real-World Robot Manipulation Policies in Simulation (2024, positive): “That is, if one policy performs better than another in real-world 1] , higher is better) for assessing the correlation between policy performances in real-world and simulation, as well as the overall quality of simulated evaluation pipelines.”
RoboCasa: Large-Scale Simulation of Everyday Tasks for Generalist Robots (2024, methodology): “RoboCasa builds upon RoboSuite [51], a modular, fast, and easy-to-use framework based in MuJoCo.”
Continual Model-Based Reinforcement Learning with Hypernetworks (2020, positive): “The environment is modified from the DoorGym environment [42] and simulated in the Surreal Robotics Suite [43].”
MimicPlay: Long-Horizon Imitation Learning by Watching Human Play (2023, methodology): “To extensively evaluate the methods with more testing trials and training seeds, we conduct an experiment in simulation LIBERO [60], which is a multitask robot manipulation benchmark based on robosuite [61] and MuJoCo [62].”
LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning (2023, positive): “Our generation pipeline is built on top of Robosuite [76], a modular manipulation simulator that offers seamless integration.”
ProcTHOR: Large-Scale Embodied AI Using Procedural Generation (2022, background): “RLBench [53], RoboSuite [133] and Sapien [124] target fine-grained manipulation.”
A Review of Safe Reinforcement Learning: Methods, Theories, and Applications (2024, positive): “3, has been developed on the basis of Robosuite [250] which is a popular robotic arm benchmark for single-agent reinforcement learning.”
Orbit: A Unified Simulation Framework for Interactive Robot Learning Environments (2023, methodology): “Frameworks that rely on CPU-vectorization ([2], [4], [10]) show an increase in throughput with the number of environments.”
iGibson, a Simulation Environment for Interactive Tasks in Large Realistic Scenes (2020, background): “However, existing simulation environments that combine physics simulation and robotic tasks often cater to a narrow set of tasks and include only clean, small-scale scenes [12, 13, 14, 15, 16, 17].”
Masked Visual Pre-training for Motor Control (2022, background): “While there exist a number of excellent benchmarks for motor control, e.g., DMC (Tassa et al., 2018), RLBench (James et al., 2020), Robosuite (Zhu et al., 2020), MetaWorld (Yu et al., 2020), they all fall short on one or more of our requirements.”