dm_control: Software and Tasks for Continuous Control
Yuval Tassa†, Saran Tunyasuvunakool†, Alistair Muldal† Yotam Doron, Piotr Trochim, Siqi Liu, Steven Bohez, Josh Merel†, Tom Erez, Timothy Lillicrap, Nicolas Heess††† Corresponding authors: {tassa, alimuldal, stunya, jsmerel, .
Abstract
The dm_control software package is a collection of Python libraries and task suites for reinforcement learning agents in an articulated-body simulation. A MuJoCo wrapper provides convenient bindings to functions and data structures. The PyMJCF and Composer libraries enable procedural model manipulation and task authoring. The Control Suite is a fixed set of tasks with standardised structure, intended to serve as performance benchmarks. The Locomotion framework provides high-level abstractions and examples of locomotion tasks. A set of configurable manipulation tasks with a robot arm and snap-together bricks is also included. dm_control is publicly available at \hrefhttps://www.github.com/deepmind/dm_controlgithub.com/deepmind/dm_control.
中文速览
连续控制与机器人研究长期缺乏一套统一、易用的仿真与任务开发平台，DeepMind 为此开源了 dm_control 软件包。该工具包以 MuJoCo 物理引擎为核心，提供了从底层物理接口（MuJoCo Python 封装）、程序化模型构建（PyMJCF）、强化学习环境 API，到高层任务编写框架（Composer）的完整工具链，并内置了三类现成任务套件：用于算法基准测试的 Control Suite、支持多种运动形态的 Locomotion 框架，以及基于机械臂的 Manipulation 操作任务。实验表明，这套工具已在 DeepMind 内部大量研究中得到验证，涵盖运动控制、模仿学习、多智能体等方向。它的重要性在于：为连续控制领域提供了类似 ALE 之于 Atari 游戏的标准化研究基础设施，让不同算法可以在统一、可复现的环境下公平比较，极大降低了新研究的上手门槛。
原文 arXiv:2006.12983;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2006.12983v2