Droids may soon scramble to aid car crash victims
Researchers have developed a robot to help stranded or injured people talk to emergency workers until human help arrives.
San Diego—A simple point of facts says the best way to spare lives when there’s a traffic accident is to get help to the scene pronto. That’s why researchers developed a robot to hit the scene immediately after a traffic accident, where its video camera and mini TV screen would allow stranded or injured people to talk to emergency workers until human help arrives.
The all-terrain wheeled rescue robot is part of a “roadside response system” developed by Mohan Trivedi at the University of California at San Diego. Busy freeways already have dedicated response teams, which clear smashed cars and other wreckage from accidents.
Trivedi’s robots would speed things up even more by alerting the emergency services earlier. “If we can get emergency services to the scene just five minutes sooner, [it] translates to 49 lives saved per year on America’s freeways,” he said. Pole-mounted cameras would monitor the road, sending video images to a roadside computer that scans them for crashes or signs of trouble, such as a vehicle pulling off to the shoulder.
When the system detects an incident, it attempts to work out what has happened by assessing whether the people involved are waving their hands for help or kneeling to change a tire. The system then alerts emergency services and deploys the droid, which trundles out of a roadside cabinet to investigate. At 1.2 meters tall, Trivedi’s robot, called Robotar, connects via a wireless link to the highway control center. Robotar’s onboard computer enables it to navigate obstacles as it travels to its destination.
To save power, Trivedi is considering dumber robots—ones that don’t carry such sophisticated computers but download instructions from a distributed computer network that gets its information from the roadside cameras. Each roadside pole will sport two digital video cameras. The first will point at a convex mirror that creates an all-around, 360° view. Software transforms the image into a flat panorama, which allows an operator to pan the entire image electronically while the camera remains fixed.
The second camera handles detailed work. It can pan, tilt, and zoom and has high enough resolution to read a car’s licence plates. Software that seamlessly overlays the pan-tilt image onto the relevant portion of the image from the first camera allows the user to effortlessly switch between the two. Alex Jones, No. 2 honcho at the California Highway Patrol, said the scheme could offer huge advantages. “Clearly there are limits to what you can do based on what you see on a camera, but the system would certainly give us a leg up in trying to ascertain the severity of the accident and making sure the right equipment and personnel are sent out.”