One of the new libraries we're working on for ROS is a web infrastructure that allows you to control the robot and various applications via a normal web browser. The web browser is a powerful interface for robotics because it is ubiquitous, especially with the availability of fully-featured web browsers on smart phones.
The web_interface stack for ROS allows you to connect to a Web-enabled ROS robot, see through its cameras, and launch applications. Beneath the hood is a Javascript library that is capable of sending and receiving ROS messages, as well as calling ROS services.
With just a couple of clicks from any web browser, you can start and calibrate a robot. We've written applications for basic capabilities like joystick control and tucking the arms of the PR2 robot. We've also written more advanced applications that let you select the locations of outlets on a map for the PR2 to plug into. We hope to see many more applications available through this interface so that users can control their robot easily with any web-connected device.
We're also developing 3D visualization capabilities based on the O3D extension that is available with upcoming versions of Firefox and Chrome. This 3D visualization environment is already being tested as a user interface for grasping objects that the robot sees.
All of these capabilities are still under active development and not recommended for use yet, but we hope that they will become useful platform capabilities in future releases.
Leave a comment