Cars or vans cruising down the highway with no driver have been in the news (see references below) of late. Google recently had one of its experimental vehicles drive itself 350 miles along California’s scenic highway; Germany’s Berlin Free University has also announced it is been successfully running tests of its MIG driverless car, and now comes news that the European Research Council has just successfully completed a test of its driverless vehicle all the way from Italy to China. What gives? Why are these cars being developed and how do they work?
Answers vary, but the overriding consensuses seems to be that cars would be much safer out on the road if they were driven by computers rather than as they are now with human beings behind the wheel. The reasoning behind this assumption is that humans are, well, human, and are thus quite prone to making not only mistakes or errors in judgment but in how they respond when the need for quick decision making is of the utmost importance. The sad fact is, human beings are ridiculously slow when it comes to processing information, making a decision, and then taking action; at least when compared with computers. Thus, the introduction of computers into the equation.
While no one is yet claiming that their vehicle will someday soon completely push humans out of the driver’s seat, the assumption remains, it’s only a matter of time before we simply tell our cars where we want to go and then let them take us there.
But how can a car/computer combination possibly take into consideration the millions of unpredictable events that occur during a simple trip to the grocery store and make correct decisions every time? The answer it turns out, is that driving a car isn’t nearly as complicated as we all might have been assuming. Consider for example, that something like ninety percent of driving is simply keeping the vehicle in its proper lane at the right speed.
To accomplish this, a car’s computer only needs to know where the edges of the road are, and some general guidelines on proper speed. For a car to know where it is relative to the road, it only needs mounted cameras connected to software that has been designed to recognize virtually every type of road edge and every type of lane divider, and then to use software that has been in use for years in robots to guide the front wheels to keep the car within certain parameters based on distance from road edges and lane dividers.
But then the dilemma of having to deal with other cars on the road comes into play. While clearly a trickier problem to handle, engineers seem to have overcome this obstacle as well by using a combination of cameras, motion detectors, carbon dioxide sensors, lasers and even new kinds of radar. The end result is a constant influx of information that tells the main computer what if anything is in the road ahead or behind, how fast it’s moving, whether it’s slowing down or speeding up, or stopping and then it makes decisions based on that information, and then sends signals to the car controls to make adjustments.
And then, so that the car will know how to get to where it’s supposed to go, the car is outfitted with GPS and mapping software that it is able to compare that information with what it “sees” in its real world natural environment; very simple.
While it is not truly clear if cars will be driving us around in the distant future, the impact of these new technologies will definitely be noticed sooner than later though as they are added on to existing cars to help us humans do a better job of driving in the near term.