The ASTRA program has had a drone and two rovers work together in Virginia, letting them arbitrate between the physical risk and the scientific gain of an unforeseen discovery.
ASTRA — Adaptive Sensing Technology for Responsive Autonomy — seeks to answer a specific question: Can a fleet of robots behave like an autonomous science team, i.e. prioritize its tasks, assess the dangers of its environment and revise its strategy along the way without waiting for instructions from Earth? The test took place on the grounds of the Virginia Tech Transportation Institute in Blacksburg, with three machines: a drone, a rover equipped with lidar mapping and a second rover equipped with a sampling arm.
The episode that the researchers remember is that of a trade-off. When the lead drone spotted an additional area of interest, it did not abandon the task at hand but paused to deal with the opportunity — exactly the type of trade-off between risk and scientific benefit expected that the fleet had to learn to weigh on its own.
The issue is not terrestrial. Communication delays make flying from Earth impractical for missions to outer planets, and lunar or Martian exploration in risk areas requires systems that can balance their own physical capabilities, mission priorities, and hazard assessment. The work is being carried out under the scientific leadership of Bethany Theiling of NASA's Goddard Space Flight Center, with Noblis, Aurora Engineering, and the University of Tulsa as partners.
Read the full article on the original source
Go to the source →