Teaching Drones to Play Tag Like a Puppy Could Deliver a Big Boost in Autonomy
Researchers at the Georgia Institute of Technology and Sandia National Laboratories have taught drones to play a game of tag via reinforcement learning — with a view to improving how autonomous systems can respond to rapid changes in environment during pursuit, evasion, or coordination.
"Reinforcement learning is just beginning to be used in commercial robotics," says project lead Spencer Jenson of the team's work. "The really nice thing about reinforcement learning is you can ignore a lot of extremely complex math — that’s probably slightly wrong anyway — and just get a best-effort algorithm that is going to be more flexible in rapidly changing scenarios. Reinforcement learning is a simulation-based guidance technique. We really like reinforcement learning because that behavior can be very hard to optimally program, but is something that reinforcement learning can just learn how to do optimally. It makes the complex simple."
The idea behind reinforcement learning...
Copyright of this story solely belongs to hackster.io. To see the full text click HERE
