{
    "info": {
        "author": "",
        "author_email": "",
        "bugtrack_url": null,
        "classifiers": [
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: Apache Software License",
            "Programming Language :: Python :: 3",
            "Topic :: Scientific/Engineering :: Artificial Intelligence"
        ],
        "description": "# **ROBEL**: **Ro**botics **Be**nchmarks for **L**earning\n\nROBEL is an open-source platform of cost-effective robots and associated\nreinforcement learning environments for benchmarking reinforcement learning in\nthe real world. It provides Gym-compliant environments that easily run in both\nsimulation (for rapid prototyping) and on real hardware. ROBEL robots are robust\nand scalable - they have facilitated over 14000 hours (as of Aug'19) of\nreal-world training with various learning-based methods. Benchmarks using\nseveral learning-based methods are provided for ease of comparison and\nextensibility. Refer to [ROBEL's webpage](http://roboticsbenchmarks.org) for\nfull details.\n\n## Robot Platforms\n\n| ![](media/cover.png \"ROBEL platforms\") |\n:----------------------:\n| ROBEL robots: D\u2019Kitty (left) and D\u2019Claw (middle and right) |\n\nROBEL introduces two robots that are modular, extensible, easy to build, and are\nlower cost compared to many existing robotics research platforms:\n\n1. **D'Claw** is a nine degree of freedom manipulation platform capable of\n   performing dexterous manipulation.\n2. **D'Kitty** is a twelve degree of freedom quadruped capable of agile\n   locomotion.\n\n## Features\n\n1. **Gym Compliant** -- ROBEL environments are fully [Gym]-compliant and can be\nused with any reinforcement learning library that interfaces with Gym\nenvironments.\n2. **Simulated backends** -- ROBEL also includes simulated equivalents of the\nintroduced benchmarks to facilitate prototyping and debugging needs.\nSimulation backend is provided by [MuJoCo].\n3. **Hardware interface** -- ROBEL is built using Dynamixel motors and\ncommunicates with the hardware device through the [DynamixelSDK].\n4. **External tracking support** -- For D'Kitty environments, external tracking\nis supported through [OpenVR] tracking.\n5. **Open-source design** -- The hardware design and build instructions are\nfully open-sourced and are available for anyone to build their own robots.\n\n[Gym]: https://gym.openai.com\n[MuJoCo]: http://www.mujoco.org\n[DynamixelSDK]: https://github.com/ROBOTIS-GIT/DynamixelSDK\n[OpenVR]: https://github.com/ValveSoftware/openvr\n\n# Getting started\n\n## 1. Hardware assembly\n\nPlease refer to the\n[Hardware Guide](http://roboticsbenchmarks.org/getting-started) for getting\nstarted with the ROBEL hardware platforms.\n\n## 2. Software installation\n\n#### 2.1 MuJoCo\n\nDownload MuJoCo Pro 2.00 from the\n[MuJoCo website](https://www.roboti.us/index.html). You should extract this\nto `~/.mujoco/mujoco200`. Ensure your MuJoCo license key is placed at\n`~/.mujoco/mjkey.txt`.\n\nAdd the following line to your `~/.bashrc` (or equivalent) in order for\n`mujoco_py` to install properly:\n\n```bash\nexport LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HOME/.mujoco/mujoco200/bin\n```\n\nRun `source ~/.bashrc` afterwards.\n\n#### 2.2 ROBEL\n\nROBEL requires Python 3.5 or higher. You can install ROBEL by running:\n\n``` bash\npip install robel\n```\n\nWe recommend doing this in a `virtualenv` or a Conda environment to avoid\ninterfering with system dependencies or existing packages.\n\nAlternatively, you can install directly from the repository by running:\n\n```bash\ngit clone --recurse-submodules https://github.com/google-research/robel.git\npip install -e robel/\n```\n\nTo run on hardware, additionally install the [DynamixelSDK] Python API:\n\n```bash\npip install git+https://github.com/ROBOTIS-GIT/DynamixelSDK.git#subdirectory=python\n```\n\n#### 2.3 Example Usage\n\n```python\nimport robel\nimport gym\n\n# Create a simulation environment for the D'Claw turn task.\nenv = gym.make('DClawTurnFixed-v0')\n\n# Create a hardware environment for the D'Claw turn task.\n# `device_path` refers to the device port of the Dynamixel USB device.\n# e.g. '/dev/ttyUSB0' for Linux, '/dev/tty.usbserial-*' for Mac OS.\nenv = gym.make('DClawTurnFixed-v0', device_path='/dev/ttyUSB0')\n\n# Reset the environent and perform a random action.\nenv.reset()\nenv.step(env.action_space.sample())\n```\n\nNot specifying the `device_path` i.e. `env = gym.make('DClawTurnFixed-v0')`\ncreates the simulated equivalent of the above hardware environment. The\nsimulated and hardware environments have the same interface.\n\nTo interactively render a simulation environment, run:\n\n```bash\npython -m robel.scripts.rollout -e DClawTurnFixed-v0 --render\n# Also try this with other tasks such as DKittyWalkFixed-v0\n```\n\n## Benchmark Tasks\n\n### D'Claw\n\n| Task      | Description                          |\n| --------- | ------------------------------------ |\n| **Pose**  | Match a set of joint positions.      |\n| **Turn**  | Turn an object to a specified angle. |\n| **Screw** | Continuously rotate an object.       |\n\n\n### D'Kitty\n\n| Task       | Description                                   |\n| ---------- | --------------------------------------------- |\n| **Stand**  | Stand upright.                                |\n| **Orient** | Align heading with a target.                  |\n| **Walk**   | Walk to a target location.                    |\n\n## Contributing\n\nWe designed ROBEL to be an easily extensible platform for new robots, tasks, and\nbenchmarks. See [`CONTRIBUTING.md`](CONTRIBUTING.md) for a guide\non how to contribute.\n\n## Citation\n\n```\n@misc{ahn2019robel,\n    title={ROBEL: Robotics Benchmarks for Learning with Low-Cost Robots},\n    author={Michael Ahn and Henry Zhu and Kristian Hartikainen and Hugo Ponte and Abhishek Gupta and Sergey Levine and Vikash Kumar},\n    year={2019},\n    eprint={1909.11639},\n    archivePrefix={arXiv},\n    primaryClass={cs.RO}\n}\n```\n\n## Disclaimer\n\nThis is not an official Google product.\n\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "",
        "keywords": "",
        "license": "Apache 2.0",
        "maintainer": "",
        "maintainer_email": "",
        "name": "robel",
        "package_url": "https://pypi.org/project/robel/",
        "platform": "",
        "project_url": "https://pypi.org/project/robel/",
        "project_urls": null,
        "release_url": "https://pypi.org/project/robel/0.1.1/",
        "requires_dist": [
            "gym (<0.14.0,>=0.11.0)",
            "mujoco-py (<2.1,>2.0)",
            "numpy (>=1.10.4)",
            "transforms3d (>=0.3.0<0.4)"
        ],
        "requires_python": ">=3.5.3",
        "summary": "Robotics reinforcement learning benchmark tasks with cost-effective robots.",
        "version": "0.1.1"
    },
    "last_serial": 5898460,
    "releases": {
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "876d5c483e0300eec53f064f0cf31496",
                    "sha256": "28585aab689db78847cb616d044a15930d929b8963892c536a5f5fa8c3248671"
                },
                "downloads": -1,
                "filename": "robel-0.1.0-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "876d5c483e0300eec53f064f0cf31496",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">=3.5.3",
                "size": 8576599,
                "upload_time": "2019-09-10T06:15:03",
                "url": "https://files.pythonhosted.org/packages/75/ea/3d8195641e239e71fc0a61aac8589b867201bd0524e1660990d1de8e4ac1/robel-0.1.0-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "113453729ce15aaa38b85c4cf711a9f1",
                    "sha256": "efa941ef0e16c4fcb2c986b4d3d07995a19889be152054a5826763181d2673cf"
                },
                "downloads": -1,
                "filename": "robel-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "113453729ce15aaa38b85c4cf711a9f1",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.5.3",
                "size": 8497163,
                "upload_time": "2019-09-10T06:15:14",
                "url": "https://files.pythonhosted.org/packages/ac/ee/965991433d662797e410dfbc4aaa79bbf92423659d4906685fb66a5e84bb/robel-0.1.0.tar.gz"
            }
        ],
        "0.1.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "0640b6cdede95c63bceffe025e5b69b5",
                    "sha256": "0914aeead52ab85ece83751667fb910648c08d4a14148edc1d6724a64c306380"
                },
                "downloads": -1,
                "filename": "robel-0.1.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "0640b6cdede95c63bceffe025e5b69b5",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">=3.5.3",
                "size": 7709326,
                "upload_time": "2019-09-27T23:08:39",
                "url": "https://files.pythonhosted.org/packages/64/59/b5190a287c06bbe354c0320450286a2498e4d54f1a958399b35a3a49a497/robel-0.1.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "3892582ce0a0f41250a726138dff02de",
                    "sha256": "00cc02d449923857036a16a3a0efb2ee9617f9facae7d50eb95048618f2bc613"
                },
                "downloads": -1,
                "filename": "robel-0.1.1.tar.gz",
                "has_sig": false,
                "md5_digest": "3892582ce0a0f41250a726138dff02de",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.5.3",
                "size": 7628865,
                "upload_time": "2019-09-27T23:08:42",
                "url": "https://files.pythonhosted.org/packages/64/1a/2324b2c81f9f5c016bbb4e59408c813faaff61939c97f73aa2919881fa40/robel-0.1.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "0640b6cdede95c63bceffe025e5b69b5",
                "sha256": "0914aeead52ab85ece83751667fb910648c08d4a14148edc1d6724a64c306380"
            },
            "downloads": -1,
            "filename": "robel-0.1.1-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "0640b6cdede95c63bceffe025e5b69b5",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": ">=3.5.3",
            "size": 7709326,
            "upload_time": "2019-09-27T23:08:39",
            "url": "https://files.pythonhosted.org/packages/64/59/b5190a287c06bbe354c0320450286a2498e4d54f1a958399b35a3a49a497/robel-0.1.1-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "3892582ce0a0f41250a726138dff02de",
                "sha256": "00cc02d449923857036a16a3a0efb2ee9617f9facae7d50eb95048618f2bc613"
            },
            "downloads": -1,
            "filename": "robel-0.1.1.tar.gz",
            "has_sig": false,
            "md5_digest": "3892582ce0a0f41250a726138dff02de",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.5.3",
            "size": 7628865,
            "upload_time": "2019-09-27T23:08:42",
            "url": "https://files.pythonhosted.org/packages/64/1a/2324b2c81f9f5c016bbb4e59408c813faaff61939c97f73aa2919881fa40/robel-0.1.1.tar.gz"
        }
    ]
}