At first glance, Carnegie Mellon University's Cadillac SUV looks like any other car on the road. But unlike other cars, this one drives itself. According to Professor Raj Rajkumar and his team at the university - the Cadillac prototype is the most advanced example of driverless car technology yet produced.
The research engineering team has been working on the autonomous car since 2000 with more than $11 million of finding from both the National Science Foundation and General Motors.
Rajkumar says driverless cars will one day be the norm, replacing the single biggest cause of traffic accidents on the road - human drivers. Human error accounts for more than 1.3 million traffic deaths globally every year and costs hundreds of billions of dollars.
"Humans get distracted," said Rajkumar. "Over 93 percent of accidents in cars happen due to human error. People are distracted, they are sleepy maybe angry, looking at something else or they may even be drunk. If we can basically take the human out of the driving equation distractions will go away and computer will not get distracted because they are not human. So, therefore we can minimize those accidents and we can slowly take those accidents towards zero," he added.
Nearly all of the major car producers globally are researching driverless car technologies with plans to phase in autonomous controls over the next 10 years. Mercedes Benz plans to offer their S-class model with autonomous capabilities like steering, braking, and parking within two years. Google's driverless car research program has logged more than 300,000 miles on roads, mostly in the U.S. state of Nevada where the first license for an autonomous vehicle was issued two years ago. Florida and California have also started working on legislation concerning driverless cars.
The Carnegie Mellon prototype is equipped with a host of sensory technologies including lasers, cameras and radar, all of which feed their data into four onboard computers, giving the vehicle a picture of its surroundings and possible dangers in real time. It also uses GPS and wireless technology to give its autonomous navigation controls a sense of direction. Rajkumar says the viability of a driverless car also depends on creating a network where cars can communicate with each other as well as with traffic controls on the road.
"It talks to the traffic lights. You see the status of the traffic lights as we go by. It is actually talking to them wirelessly," Rajukumar said during a test drive of the prototype.
He says the car's sense of awareness is far superior to a human driver's - especially is dangerous conditions.
"Suppose you are driving at night and the streetlights are off and the headlights start blinking and they die just to take an extreme example. We as humans can not perceive anymore but if the car is outfitted with a thermal imaging camera, a heat-sensing device it can actually see if there is a human or an animal in front of the car. The computer can actually detect that and stop the car. We humans can't perceive the obstacle but the computer has," he said.
But Rajkumar says his team is still working on proving that their driverless car can reliably handle all of the variables and complexities involved in what seems like ordinary drive down the road.
He says he is confident that by the year 2020 - his autonomous car will be ready. But he says getting people comfortable with the idea of letting their cars drive them around may take a little longer.