A team from South Korea finished top in the DARPA-run robotics challenge held in LA over the weekend. The bot-based battle saw Team Kaist’s DRC-Hubo robot beating 22 rivals in a series of challenges, a victory which scored it $2 million in prize money. Florida-based Team IHMC Robotics picked up the $1-million second prize with its Running Man robot, while the $500,000 third prize went to Tartan Rescue of Pittsburgh with a bot called CHIMP.
The two-day event saw the robots taking on a demanding course involving eight tasks connected with disaster response. The challenges including walking over rubble, tripping circuit breakers, turning valves, and climbing stairs. All eight tasks had to be successfully completed within 60 minutes for a team to have a chance of bagging the top prize. Hubo (pictured) managed to complete all eight tasks in 44mins 28secs, while second-placed Running Man came in six minutes later.
The event – a contest focusing on technology developed to aid emergency personnel in disaster environments – was set up following the earthquake, tsunami, and nuclear meltdown that hit northeast Japan in 2011.
The Fukushima nuclear plant suffered a series of explosions in the incident, and with high radiation levels inside the plant too high for humans, robots were sent in to deal with the damage. But with debris spread throughout the wrecked facility and spaces too tight to comfortably negotiate, many of the bots weren’t up to the job.
DARPA says its international event, which also included several elimination rounds over the last two years, has helped to accelerate advancements in robot technology and “hasten the day when robots have sufficient dexterity and robustness to enter areas too dangerous for humans and mitigate the impacts of natural or man-made disasters.”
You can marvel at Hubo performing some of his moves in a video shot by his creators shortly before the DARPA Robotics Challenge event. One smooth operator, Hubo can be seen effortlessly making his way over rubble, walking up steps, and even driving a vehicle.