Wired writes that: “The next generation of gearheads won’t obsess over horsepower and torque; they’ll focus on things like radar range, communication latency, and pixel resolution” in an article that examines the current state of cars that drive themselves. Some of the things they found include:

• Premium cars today run 100 million lines of computer code. This is more than Boeing’s new 787 Dreamliner Aircraft
• Google compares teaching cars to drive themselves to mining the web so that people can find the pages they’re looking for when they search. For example just as Google’s algorithms must go through large amounts of data on the web, they must also analyze road data to decide what certain objects are and how they can be navigated around.
• Mercedes has also been working to develop the technology. They note that cars can already do a lot of automatic work. Their cars:
  • Monitor 70 points of information about baseline data such as steering wheel movements and indicator use. When performance is significantly different from the baseline the car assumes that the driver is fatigued and suggests getting a cup of coffee.
  • Automatic wiper activation in rain.
  • Automatic headlight illumination.
  • Notifications in the sideview mirrors if another car is in the driver’s blind spot.
  • Lane drift detection
  • Automatic breaking systems.
• Challenges still remain. Cars must learn to drive on roads without lane markings, and must adapt to local driving cultures.

To read more about how the cars of the future will be like the horses of old, why these changes won’t make us feel uncomfortable, and some of the moral and ethical challenges that the new technology poses click here.

Source: Wired

Via: Marginal Revolution