Main Page

From Brown University Robotics

The Brown University Robotics Group strives to realize robots and autonomous systems that are effective collaborators for humans. Towards this end, we explore problems in human-robot interaction, robot learning, robot perception, autonomous control, dexterous manipulation, and game development. Our approach to these problems often uses techniques from machine learning. Brown Robotics aims to enable reproducibility and interoperability of our research and development work within robotics and across society. Brown Robotics is an active contributor to the ROS effort through our brown-ros-pkg repository.

In collaboration with Bosch Research, we are starting to get the PR2 Remote Lab up-and-running. Check out the introductory video for a look at some of the Remote Lab features.

The PR2 Remote Lab is a web-centered remote robot laboratory comprising a PR2 and the hardware/software infrastructure necessary for making it available for public Internet use. This remote laboratory is designed to allow a wide group of researchers to conduct experiments on a PR2 robot. The infrastructure used to create this lab is open source and can be used to interact with any PR2 robot, and any robot running ROS. This site provides documentation to allow users to begin to interact with the robot in the remote lab as well as to use this documentation with any PR2 robot.

Much of the software we develop is available through the Brown Ros Package collection, which features:

• irobot_create_2_1: A driver for the iRobot Create, including a lightweight Python binding that exposes most of the Open Interface functionality.
• gscam: A ROS webcam capture node which leverages Gstreamer, making it compatible with almost every Linux camera and video system available. In addition Gstreamer's software video processing can be used to emulate advanced features (e.g. white-balancing) even for cameras that don't have the appropriate v4l hardware support.
• ar_recog: A ROS image processing node which leverages ARToolKit to recognize AR (Augmented Reality) tags and their location/transformation in image-space.
• teleop_twist_keyboard: A simple keyboard-based teleop interface inspired by teleop_base. It's only dependencies are on the necessary geometry messages.
• position_tracker: A ROS node allowing for simple odometry correction.
• pr2_keyboard_teleoperator: ROS teleoperation for most functions of the PR2.
• aldebaran_nao: Basic ROS driver for the Aldebaran Nao

For older information, please refer to the previous Brown Robotics site.