Project Tango was just announced by Google’s ATAP group (http://www.google.com/atap/projecttango/). The project developed 3D localization and mapping capable of running in real-time on a phone. We expect the technology developed as part of Project Tango to be transformative for the robotics industry and are proud to have been a part of the process.
OSRF has been a partner with ATAP on the Tango project since May 2013. Our role was to bring our open source ROS (http://www.ros.org) software and expertise to bear on the problems that Tango is tackling. While Tango isn’t strictly a robotics project, there is a lot of similarity between their needs and what we have encountered over the years working with robots.
Among our contributions were developer tools for debugging, data logging, and data visualization. Project Tango used popular ROS tools, including rviz and rosbag, for development and debugging on a second screen, such as a desktop or laptop, both for live and recorded data.
Somewhat surprisingly, we also helped out with managing the complex code base that became Tango. With a large team of geographically distributed developers working at a furious pace, it’s not easy to keep track of the software in a project like this. Fortunately, we’re intimately familiar with this challenge from the past six years of ROS development. Project Tango was able to leverage the catkin build system developed for the ROS ecosystem to bring together the work from their many contributors.
In addition to integrating new code, catkin was used to integrate many existing libraries including familiar ROS dependencies such as OctoMap, OpenCV, and Eigen. To support both on-device and desktop debugging, catkin was used to enable compiling for Android devices as well as desktop targets from the same source tree.
Our work on Project Tango resulted in new features, bug fixes, and polish being added to many of our core tools, all of which have been made available to our community.