Ann Arbor, Michigan– Teaming with the University of Michigan, SoarTech was recently awarded first place and $750,000 at the first Multi Autonomous Ground-robotic International Challenge (MAGIC 2010). SoarTech’s real-time 3D interface featuring SAGE (Situation, Actions, Goals, Environment) impressed the judges and helped Team Michigan take first in the competition to develop a team of robots that could operate autonomously in urban combat zones.
The nearly two year long, international competition was sponsored by the U.S. Army Research, Development and Engineering Command’s Tank Automotive Research, Development and Engineering Center (TARDEC) and Australia’s Defense Science and Technology Organization. Participants were challenged to develop a team of autonomous ground robots capable of communicating and completing tasks while operating with minimal operator intervention. Team Michigan used 14 robots, more than any other finalist, to effectively map areas of an urban environment and neutralize observed threats. The robots coordinated their actions and passed information back to a ground station manned by two human operators. One operator was in charge of assigning tasks while the other managed the classification and engagement of objects of interest. The SAGE interface provided the operators with the necessary information to direct and control the system.
Team Michigan’s use of SoarTech’s SAGE interface demonstrated the capability for numerous robots to be controlled by only two human operators. The key feature of the SAGE interface is its ability to automatically detect and highlight events that are important to the user. Underneath the display is event detection logic that autonomously detects events of interest for an operator, such as a civilian close to a dangerous object or a robot about to execute neutralization. SAGE prioritizes and tracks each of these events, reducing clutter by combining events that are related. For the highest priority events SAGE automatically creates a 3D first-person view of the event to provide critical spatial and status context to the operator. In the future, SoarTech will enhance the event detection logic with deeper models of the robots’ tasks, the operators’ goals, the situation, and the environment in order to do real-time, goal-based event stream analysis. These enhancements will allow the SAGE interface to better tailor the automated alerts to the specific mission and situation. Additionally, the SAGE interface will be able to take over more routine robot control tasks further reducing the operator’s cognitive load.
Competition judges found the SAGE interface to be an important component in helping the operators as they managed multiple autonomous robots performing tasks in a dynamic environment.
Team Michigan was one of five finalists selected out of an initial 23 to compete in the final stage of MAGIC 2010. A team of U-M students, led by Professor Edwin Olson of the College of Engineering, and SoarTech employees traveled to Adelaide, Australia to compete at the Royal Fairgrounds, November 7-12.
The team from the University of Pennsylvania came in second, winning a $250,000 grant. Team Reconnaissance and Autonomy for Small robots from Gaithersburg, Maryland came in third and was awarded $100,000.
About SoarTech
SoarTech makes the vision of true autonomy a reality every day. By combining artificial intelligence with cognitive and social psychology SoarTech builds intelligent systems for defense, government and commercial applications that bring human-level intelligence and decision making to robotic systems. Their goal is to enhance human operations and abilities by creating thinking partners. Partners that work to make humans more prepared, more informed and more capable.
To learn more about SAGE and SoarTech, visit www.soartech.com.
Media Contact
Andrew Dallas
SoarTech
734-327-8000 x214
.(JavaScript must be enabled to view this email address)
