Considering the number of traffic accidents involving driver error, it seems obvious that many lives would be saved if cars could drive themselves. Additional benefits of autonomous vehicles include less traffic congestion and better fuel economy/lower emissions.
The enabling technology is close at hand. Google’s fleet of driverless cars has traveled nearly 200,000 miles, DARPA (Defense Advanced Research Projects Agency) has sponsored three Grand Challenge events, and automotive electronics applications such as adaptive cruise control, lane keeping assist, and parallel park assist are increasingly common.
So near and yet so far, as an article by John Markoff in the January 23 New York Times makes clear. Speakers at a symposium sponsored by the Santa Clara Law Review (http://lawreview.scu.edu/page.cfm/events) suggested that questions of legal liability, privacy, and insurance regulation could pose more problems than the technology.
Not that the technology is a slam dunk. Among the problems remaining to be solved are how a driverless vehicle can recognize a safety officer motioning the vehicle to pull over or to proceed in a different direction than the vehicle intends. Or how a driverless vehicle programmed to obey all traffic regulations will fare against human drivers who cut metaphorical corners. Sven A. Beiker, executive director of the Center for Automotive Research at Stanford University, suggests that completely autonomous vehicles may be operating on limited roads in 20 years. Autonomous driving presents a huge ongoing reliability challenge for automotive electronics engineers
It’s good to know that attention is being paid to liability, privacy, insurance, and other issues in addition to technology. It’s encouraging to think that in as little as 20 years driverless vehicles will be a reality albeit in limited form. In the meantime we can look forward to more quasi-autonomous systems that give drivers the option of letting the car take over.