In a perfect world, cars and trucks would never turn left. Left turns waste time, gas and are dangerous for drivers, oncoming traffic and crossing pedestrians. So imagine teaching a machine to turn left — in Boston’s infamous traffic, no less? A driverless car has to read human psychology, subtle signals from other drivers and pedestrians to navigate one of the hardest maneuvers on the road.
As engineers race to build self-driving cars, they’ve found that making them safely turn left is one of their toughest problems. “I see a lot of challenges every day, and left turns is near the top of the list,” said John Leonard, a professor at the Massachusetts Institute of Technology who specializes in self-driving vehicles.
Left turns are so tough because they involve psychology as well as technology. Drivers and pedestrians read subtle signals from each other as they approach an intersection. We’ve come to learn how to read those signals to make pretty good guesses about when it is safe to turn left in a variety of traffic conditions.
“Human beings are remarkably good at that, because we’re social beings,” said Gill Pratt, chief executive of the Toyota Research Institute, an organization founded by Japanese carmaker Toyota Corp. to solve critical problems in automotive and robotic technology.
Read the source article at the Boston Globe.