Random Fact: There has been an average of 10 to 11 million motor vehicle accidents annually in the US since 2004 (Source: U.S. Census Bureau)
Another Random Fact: Only ~12% of a car’s energy use goes toward providing momentum/moving the passenger (Source: Hofstra)
While both facts/issues have garnered the interests of academia and industry folks, some of the solutions proposed thus far have been counter-intuitive, but expectantly diverse. Academia seems preoccupied with creating smart-phone applications and providing the driver with more responsibilities and distractions, while on the other hand, private enterprises are leaning more toward car automation and yielding less responsibilities to the driver. In a not-too-recent article, UC Berkeley and IBM announced plans of a partnership to create a smart-phone app that would be the equivalent of a “prediction” model for daily traffic in order to combat congestion and fuel inefficiency, given a driver’s GPS data history. And on the East Coast, researchers at MIT and Princeton were reported to having developed a smart-phone traffic app that provides real-time traffic signals in advance to drivers for the sake of improving fuel efficiency. The catch: its a crowd-sourcing app that relies on high traffic activity in order to be effective.
The private enterprise approach has been the more costlier model but, in the long run, it proves to be more effective in reducing motor vehicle accidents and improving fuel efficiency. I remember, a while back, reading an article about a Google project aimed at fully automating the car driving experience. Although the project is still far from being market-ready, the effort is definitely a step forward. However, with that being said, I do not think a human driver will ever be fully replaced when it comes to the ubiquitous automobile given the multitude of changing environmental variables on any given route and on any given day that a computer may not simply be capable of accurately assessing 100% each time. I liken the scenario to GPS-guided smart bombs that can become error-prone due to electronic noise. In addition to its autonomous vehicle project, Google has recently partnered with Ford Motors on a project similar to UC Berkeley and IBM’s “prediction model”. The only difference is that the Google and Ford project would integrate a car’s computer with the cloud, thus providing the car with real-time decision-making abilities, instead of relying on a smart-phone app.
In my opinion, the private enterprise is correct in focusing on further automating the car driving experience.