 |
Student Robotics
Competition October 14, 2007 Orlando, FL |
| Home | Rules | Qualification | On-Site Competition |
Welcome to the Tapia Robotics Competition 2007!
The Tapia Robotics Competition 2007 is focused around a search and rescue task as means to foster participation in robotics and computing by diverse populations of students. Technically, the competition involves programming a mobile robot to autonomously search for and visit a number of markered, but unknown, locations (or survivors) in a given environment. Upon indentifying and reaching each survivor, your robot should "call for help." Each valid call for help will be rewarded with an increased score. The qualification stage will be conducted in a simulation environment. All participants that meet the minimum criterion will be invited to compete in the the next stage of the competition which will be conducted at the conference site. The onsite competition will involve physical robots solving the search task in head-to-head competition. Although the environment is a simplified mockup of an actual search and rescue scenario, the basic principles of autonomous robot control remain the same.
The competition is a staged event, starting from qualification with simulated robots up to actual competition with the iRobot® Create™ mobile robot. Tapia Robotics supports a standardized robot middleware package (Player/Stage/Gazebo) for robot programming. Player is a client/server system that manages robot control similar to how web browsers and servers manage webpages. Each robot runs a robot server that exposes access to its sensors and motors through an Internet (TCP/IP) port. Students write robot clients that connect to a robot server to access its data, decide the robot should do, and send appropriate control commands to the robot. Robot clients can be written in any computer language supported by Player/Stage, such as C/C++, Java, and Python.
 Simulated robot with range and video sensors |
The
interfaces to the robot's sensors and motors are abstracted by
proxies that are device-independent descriptions of robot
hardware. Examples of these proxies (called interfaces in Player terminology) include position for
driving the robot's motors and reading odometry, camera for
getting images from the camera, and blobfinder for finding objects
of solid colors from camera images. This concept is similar to the
hardware abstraction layer in modern operating systems that
allow for plug-and-play devices. Thus, Player/Stage
controllers are not specific to any particular hardware and can be
easily ported between different types of robots, such as a Segway,
iRobot Roomba, ActivMedia Pioneer, etc.
For qualification, students will develop Player/Stage controllers for a physically simulated environment. The Gazebo 3D simulation platform enables Player/Stage robot clients to control simulated versions of ActivMedia Pioneer2DX. Gazebo utilizes the widely-used Open Dynamics Engine for rigid body physics. |
Students will then port and modify their qualification controller a real mobile robot for participation in the actual competition. Teams are free to choose their mobile robot platform. An example of a suitable platform is the Smurv created by the Brown Robotics Group (Robotics, Learning, and Autonomy at Brown). The Smurv has an onboard mini-ITX computer, with a 1.2GHz Pentium class processor running SLAX Linux, that is powered by and drives an iRobot Create. Specifications for this platform along with suggestions for other alternatives will be given after initial qualification rounds.
|  |
| Smurv Hardware Platform (front and back) | |