DARPA's Virtual Robotics
When disasters strike, the danger persists long after the storm, earthquake or tornado has gone. In addition to those directly impacted by such events, rescue workers and cleanup crews face considerable risks during and after a natural disaster.
For example, during the Fukushima nuclear facility meltdown following the devastating tsunami in Japan, many technicians were exposed to deadly levels of radiation while trying to contain the crisis. Emergency and disaster relief crews also face considerable hazards from unstable structures, panicked citizens and ongoing forces of nature.
The DARPA Virtual Robotics Challenge, which began on June 17, 2013, hopes to spur the development of adaptable robots designed to remove humans from such life-threatening scenarios. Specifically, DARPA seeks to prove the possibility of three robotic capabilities:
-
- Compatibility with environments engineered for humans (even if they are degraded)
- Ability to use a diverse assortment of tools engineered for humans (from screwdrivers to vehicles)
- Ability to be supervised by humans who have had little to no robotics training
According to DARPA, the third objective is the most critical, as supervised autonomy “allows simple tasks to be performed by the robot without full-time operator intervention.” The first phase of the challenge seeks to address the issues of robot-operator perception, manipulation and locomotion.
Teams have thus been tasked to control a simulated humanoid robot to accomplish a set of physical challenges, using OSRF’s Gazebo simulator and the Robot Operating System (ROS) language. In the virtual environment, DARPA simulates the hectic conditions of a real natural disaster by limiting teams’ knowledge of the scenario and artificially introducing communication issues between the team and the virtual robot. Teams are also limited to the first person view of the robot, to more closely replicate a real life scenario.
Atlas robot
Evaluated on their ability to complete a variety of physical tasks, the top teams will be awarded with an Atlas robot and funding to assist them in the next DARPA challenge, to take place in December.
The winning teams will be announced on June 27.
