Google self-driving car a big data project
Asked what inspired him to build the Google self-driving car, Sebastian Thrun mentions being at the DARPA Grand Challenge in 2003, a long distance competition for driverless cars, funded by the United States Department of Defense. Witnessing how these cars went up in flames after only a mile, he was convinced he could do better. When the US government proposed a million dollars for a research team able to build a self-driving car that could cross the desert, he took the challenge. He put together a team and spent a year in the front seat of a car writing computer code. Their car went on to win the Grand Challenge in 2005, completing the whole course.
Since then a lot has changed. The invention of visual feedback using radar, cameras and sen-sors helped overcome the shortcomings of GPS, which could not keep a car centered on a lane for very long. And it became apparent that a driverless car had larger implications than just mili-tary use. On the one hand a “smart” car could make driving safer and reduce the millions of deaths caused by car accidents each year. On the other hand, automotive transport still is a billion-dollar industry and the second biggest expense in US households.
Thrun, at the time a professor of computer science at Stanford University, was then approached by Google. As he states in a video interview, his move from Stanford to Google was not mainly about the economic potential a self-driving car seemed to have; there was also another reason: “It’s not about the hardware itself, the car itself,” he says. “It’s really about the data. And what better place is there than Google with its large data centers and information processing?”
As it scans and analyzes its environment, capturing everything that it sees moving, the Google car gathers nearly 1 gigabyte of data per second, blending it all together to make its decisions while driving. According to analysts, this will be attracting the attention of regulators because a lot of privacy, cybersafety and security issues are involved. "It's extremely valuable data," IDC analyst Sheila Brennan says. "I can't argue that point. That data will be worth a lot, and it's still not clear, again, how the consumer will play out."
Big data might therefore even turn out to be a “major roadblock” to getting these cars on the road. Despite all the excitement over the advancements in autonomous vehicle technology, too many questions remain to be answered before the driverless car becomes a common sight.