Sarah Osentoski of Brown's RLAB recently announced a beta version of a ROS to RL-Glue bridge for reinforcement learning

Brown is pleased to announce our beta version of rosglue. rosglue is a bridge between ROS and RL-Glue, a standard reinforcement learning (RL) framework.

rosglue is designed to enable RL researchers and roboticists work together rather than having to reimplement existing methods in both fields. A goal of rosglue is to allow ROS users to use RL algorithms provided by RL researchers and, likewise, to allow RL researchers to more easily use robots running ROS as a learning environment. rosglue allows a robot running ROS to become an RL-Glue environment allowing RL-Glue compatible agents to control the robot. A high level visualization of the framework can be seen here.

rosglue uses a yaml configuration file to specify the topics and services and the learning problem. rosglue automatically subscribes to the topics and services specified in the file. rosglue sends actions selected to the RL-Glue to the robot using the appropriate topic or service. It then creates observations from specified topics for the RL-Glue agent.

rosglue is currently available for download from the brown-ros-pkg repository via:

svn co https://brown-ros-pkg.googlecode.com/svn/trunk/experimental/rlrobot/rosglue rosglue

and preliminary documentation can be found here:

http://code.google.com/p/brown-ros-pkg/wiki/rosglue

Robot Learning and Autonomy @ Brown (RLAB)

The CCNY Robotics Lab, which was recently featured in this CityFlyer blog post, has just announced the release of two packages for laser scan registration.

Dear ROS-Users,

The CCNY Robotics Lab is pleased to announce the release of two packages for laser scan registration. canonical_scan_matcher is a wrapper around Andrea Censi's "Canonical Scan Matcher" [1]. polar_scan_matching is a wrapper around Albert Diosi's "Polar Scan Matching" [2].

Both packages estimate the displacement of a robot by comparing consecutive Laser Scan messages. They can be used without providing any estimate for the displacement of the robot between the scans. In this way, they can serve as an odometric estimate for robots that don't have any other odometric system. Alternatively, a displacement estimate can be provided as input to the scan matchers, in the form of an Imu message or a tf transform, in order to produce better (or faster) scan matching results.

While the two scan matchers use different algorithms and parameters, the ROS wrappers are identical in terms of topics/frames/tf's, making the two packages interchangeable.

Documentation and usage instructions can be found at the respective wiki pages:

• canonical_scan_matcher
• polar_scan_matcher

As usual, we have provided a small demo bag file with laser data and a launch file that can be used to view the packages in action. Each wiki page also has a video of what the output of the demo should look like.

We hope you find the scan matchers useful, and we extend our thanks to the authors of the original implementations.

Ivan Dryanovski
William Morris
The CCNY Robotics Lab

[1] A. Censi, "An ICP variant using a point-to-line metric" Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2008
[2] A. Diosi and L. Kleeman, "Laser Scan Matching in Polar Coordinates with Application to SLAM " Proceedings of 2005 IEEE/RSJ International Conference on Intelligent Robots and Systems, August, 2005, Edmonton, Canada

Skybotix is offering their CoaX helicopter complete with basic ROS setup so customers can use ROS right out of the box.

The CoaX helicopter is a micro UAV targeted at the research and educational markets. The small 320g helicopter includes an IMU, a downward-looking and three optional sideward-looking sonars, pressure sensor, color camera, and Bluetooth, XBee, or WiFi communication. In addition to two DSPs (dsPIC33), the CoaX has an optional Gumstix Overo computer that can run ROS. You can see more of the specs on their hardware wiki page.

Skybotix fully supports open source with the CoaX. The CoaX API, including low-level firmware and controller, is available open source under a GNU LGPL license. Their Gumstix Overo setup comes with a basic ROS installation. They include a ROS publisher for the CoaX state, a demo application for transmitting video data, and a GUI for visualizing both. Although the CoaX comes with minimal additional ROS libraries, there is a growing community of micro-UAV developers using ROS, including the micro-UAV-focused ccny-ros-pkg repository.

The CoaX was developed in collaboration with ETH Zurich. The Skybotix Youtube channel has videos of ETH Zurich student projects. Skybotix released recently a speed module for CoaX based on optical sensor, which enables indoor speed control as well as indoor hovering (video).

DARPA is having a contest to name their new robot for the ARM program. "The ARM Robot" has two Barrett WAM arms, BarrettHands, 6-axis force torque sensors at the wrist, and pan-tilt head. For sensors, it has a color camera, SwissRanger depth camera, stereo camera, and microphone.

The final software architecture and APIs have not been released yet, but the FAQ notes:

The software architecture is TBD, but is leaning toward a nodal software architecture using a tool such as Robotic Operating System (ROS).

The software track for the ARM program currently includes Carnegie Mellon University, HRL Laboratories, iRobot, NASA-Jet Propulsion Laboratory, SRI International and University of Southern California. It would certainly be a great boost for the ROS community to have more common platforms to develop and share the latest perception and manipulation techniques.

Below is a video from Dr. Motilal Agrawal of SRI (via Hizook) showing it in action. Dr. Agrawal and SRI are looking for Ph.D/Masters students with experience in robotics, ROS, and OpenCV. Want a job?

In addition to WowWee drivers and OpenCV tutorials, I Heart Robotics has just released a great Ubuntu panel tool called RIND, which stands for Robot/ROS Status Indicator. You can use it to manage your local roscore as well get information on ROS nodes and topics. Checkout the documentation or read the announcement for more information.

The CityFlyer project at the CCNY Robotics and Intelligent Systems Lab is using Ascending Technologies Pelican and Hummingbird Quadrotor helicopters to do research in 3D mapping and navigation. The Ascending Technologies platform provides a 1.6Ghz Intel Atom processor, 500 gram payload, GPS, and barometric altimeter. The CityFlyer add several sensors, including a Hokuyo URG-04LX and IMU. The Hokuyo URG has been modified to double as a laser height estimator. The CityFlyer project is able to combine data from these sensors to do indoor SLAM using GMapping.

The CityFlyer project has also created an RGB-D sensor by combining data from a SwissRanger 4000 and Logitech Webcam. They use this to build 3D maps for indoor environments using a 3D Multi-Volume Occupancy Grid (MVOG). Their MVOG technique is described in their RGB-D 2010 paper and more videos are here and here. Although the full sensor package exceeds the payload of the quadrotor, they anticipate that advances in RGB-D will make these techniques feasible for micro UAVs.

CCNY has released a variety of drivers, libraries and tools to support the ROS community. These include drivers and tools for the AscTec platform, libraries for dealing with aerially mounted laser rangefinders, a New College Dataset parser, and libraries for using AR tags with ROS.

CCNY has also developed a "Ground Station" application that acts as a virtual cockpit for visualizing telemetry data from an AscTec quadrotor. It is also able to overlay GPS data on an outdoor map to visualize the UAV's tracks. I Heart Robotics has a great writeup on Ground Station, and you can also checkout the documentation on ROS.org.

The ccny-ros-pkg is an excellent resource for the ROS community with complete documentation on a variety of packages, including videos that demonstrate these packages in use.

Bag files for the video above can be downloaded here (elevator_2010-08*).

ROS C Turtle Update 1 has been released. The list of updates includes:

• pr2_tabletop_manipulation_apps 0.2.1
• point_cloud_perception 0.2.8
• topological_navigation 0.1.2
• simulator_gazebo 1.2.3
• pr2_simulator 1.2.1
• sql_database 0.2.1
• pr2_object_manipulation 0.2.1
• physics_ode 1.2.2
• object_manipulation 0.2.2
• tabletop_object_perception 0.2.1

Qbo is a personal, open-source robot being developed by Thecorpora. Francisco Paz started the Qbo project five years ago to address the need for a low cost, open-source robot to enable the ordinary consumer to enter the robotics and the artificial intelligence world.

A couple months ago, Thecorpora decided to switch their software development to ROS and have now acheived "99.9%" integration. You can watch the video below of Qbo's head servos being controlled by the ROS Wiimote drivers, as well as this video of the Wiimote controlling Qbo's wheels. Their use of the ROS joystick drivers means that any of the supported joysticks can be used with Qbo, including the PS3 joystick and generic Linux joysticks.

Qbo's many other sensors are also integrated with ROS, which means that they can be used with higher-level ROS libraries. This includes the four ultrasonic sensors as well as Qbo's stereo webcams. They have already integrated the stereo and odometry data with OpenCV in order to provide SLAM capabilities (described below).

It's really exciting to see an open-source robot building and extending upon ROS. From their latest status update, it sounds like things are getting close to done, including a nice GUI that lets even novice users interact with the robot.

Qbo SLAM algorithm:

The algorithm can be divided into three different parts:

The first task is to calculate the movement of the robot. To do that we use the driver for our robot that sends an Odometry message.

The second task is to detect natural features in the images and estimate their positions in a three dimensional space. The algorithm used to detect the features is the GoodFeaturesToTrackDetector function from OpenCV. Then we extract SURF descriptors of those features and match them with the BruteForceMatcher algorithm, also from OpenCV.

We also track the points matched with the sparse iterative version of the Lucas-Kanade optical flow in pyramids and avoid looking for new features in places where we are already tracking another feature.

We take the images to this node from image messages synchronized and send a PointCloud message with the position of the features, their covariance in the three coordinates, and the SURF descriptor of the features.

The third task is to implement an Extended Kalman Filter and a data association algorithm based in the mahalanobis distance from the CloudPoint seen from the robot and the CloudPoint of the map. To do that we read the Odometry and PointCloud messages and we send also an Odometry message and a PointCloud message with the position of the robot and the features included in the map as an output.

Lego Mindstorms NXT is a low-cost programmable robotics kit that is used in education and by hobbyits throughout the world. One of the most visible NXT events is First Lego League. The developers of foote-ros-pkg have developed a bridge that connects NXT with ROS, allowing NXT users to leverage all the ROS tools and capabilities.

The NXT-ROS software stack provides many useful tools to interface NXT robots with ROS. Currently NXT users can take robot models created with Lego Digital Designer, and automatically convert them into robot models compatible with ROS. The converted robot model can be visualized in rviz, and in the future we hope to add simulation capabilities in Gazebo, our 3D simulator. The bridge between NXT and ROS creates a ROS topic for each motor and sensor of the NXT robot.

Once a robot is connected to ROS, you can start running applications such as the base controller, wheel odometry, keyboard/joystick teleoperation, and even assisted teleoperation using the ROS navigation stack. The NXT-ROS software stack includes a number of example robot models for users to play with and to get a feel for using NXT with ROS.

This new NXT-ROS software stack provides NXT users access to the open-source ROS community. NXT users now have access to state of the art open source robotics libraries available on ros.org.

Please see the nxt page on the ROS wiki for documentation, demos, and more. The developers would like to thank the nxt-python project for support and development.

One of the new features in ROS C Turtle was a critical component of our recent "hackathons." When fetching a drink out of a refrigerator, for example, a robot has to perform numerous tasks such as grasping a handle, opening a door, and scanning for drinks. These tasks have to be carefully orchestrated to deal with unexpected conditions and errors. We've previously used complex task-planning systems to orchestrate these actions, but our developers and researchers needed something more rapid for prototyping robot behaviors.

One of our interns came up with an answer. SMACH ("State MACHine", pronounced "smash") is a task-specification and coordination architecture that was developed by Jonathan Bohren as part of his second internship here at Willow Garage. Jonathan came to us from the GRASP Lab at University of Pennsylvania and is now headed off to the Laboratory for Computational Sensing and Robotics (LCSR) at Johns Hopkins.  During his extended stay here, SMACH was used in a variety of PR2 projects.

SMACH was first used in the rewrite of our plugging and doors code, then further refined during our billiards, cart-pushing, and drink-fetching hackathons. In all of these projects, the ability to code these behaviors quickly was critical, as was the ability to create more robust behaviors for dealing with failure.

SMACH is a ROS-independent Python library, so it can be used with and without ROS infrastructure. It comes with important developer tools like a visualizer for the current SMACH plan and introspection tools to monitor the internal state and data flow. There are already many SMACH tutorials that can be found on the ROS wiki, and we hope to see SMACH used to produce many more cool robotics apps!