Self-driving vehicles are making the news a lot these days, whether it’s with Google’s autonomous little egg, or the crashes that are being attributed to Tesla’s system.
But you might not realize there are a number of more mainstream models that can find their own way down the highway, even if they’re not advertised as such. They’re not autonomous, but they use technologies that are the building blocks for self-drivers.
Vehicles that can currently do this, including models from manufacturers such as Acura, BMW, Infiniti, Mercedes-Benz and Volvo, among others, use adaptive cruise control to maintain their distance from the vehicles in front, and lane-keeping assist to stay between the lines on the road. For safety and legal reasons, they will only do this for a few seconds before requiring you to take control again.
“The driver still needs to have his eyes on the road and hands on the steering wheel,” says Per Lenhoff, senior manager of Volvo Cars Safety Centre in Sweden. “It’s a step towards autonomous driving, but it’s a support system so you are still in charge.”
RELATED
Motor Mouth: Who will make decisions for self-driving cars?
Exactly how each technology works can vary between manufacturers, but the basic principles are the same. On Volvo’s system, called Pilot Assist, a radar system mounted behind the windshield determines how far ahead other vehicles are; once the driver sets the cruise control speed, the system maintains a pre-set distance behind the car in front. If there’s enough room, the Volvo will keep the selected speed. But if that car slows down or someone else cuts in, the system decelerates to maintain the distance. Once that other car speeds up or leaves the lane, Volvo’s system accelerates back up until it reaches the set speed.
The vehicle can come to a complete stop by itself if necessary, and if it sits for less than three seconds, it can start moving by itself again when the car ahead starts to drive.
Per Lenhoff, senior manager of Volvo Cars Safety Centre in Sweden
Drivers can use adaptive cruise control by itself, but adding lane-keeping assist gives the vehicle the tools to drive forward on its own. With Volvo’s system, a windshield-mounted camera figures out where the lines are on the road. “It activates when it sees both lane markers,” Lenhoff says. “It needs both to get its bearings, but if one (line) disappears after that, it can go on one.”
The system keeps the car between the lines by automatically adjusting the electrically assisted power steering. On its own, it can handle slight curves but isn’t meant for twisty roads. It’s primarily designed to provide subtle assistance to a driver who already has both hands on the wheel.
Numerous manufacturers use camera-based systems to detect lane markings, whether simply as a warning system when a driver is drifting across the line, or as a lane-keeping system to help keep the vehicle inside the lane. Depending on the automaker and how aggressive the correction is meant to be, this can be achieved by controlling the steering or by lightly applying the brakes on one side to nudge the vehicle back over. The system doesn’t react if the driver uses the turn signal, indicating an intended lane change.
RELATED
Motor Mouth: Here's how to confuse a self-driving car
Volvo intends Pilot Assist to be just a driver assistance program, Lenhoff says. “It’s optimized for freeways with few turns,” he says. “We want the driver to be in control, but it keeps the distance to the cars in front of you. A machine never gets tired, while you might get tired and put your eyes elsewhere. This is a step in the journey toward autonomous cars, but it’s still a supervised car.”
While almost all automakers are working on versions of this — I’ve even ridden in a Freightliner tractor-trailer prototype that’s capable of driving by itself on a freeway for countless miles — there are still some serious hurdles to overcome before we all jump in our cars, set up a movie and enjoy a completely hands-off trip.
The driver of a prototype Acura RLX sedan places his hands on his knees during a driving demonstration in Detroit, Tuesday, Sept. 9, 2014. The car has cameras that monitor lane marking and multiple radar sensors on the front and sides. On top is a beacon that uses laser beams to continually scan the car’s surroundings, similar to self-driving prototypes already introduced by Google, Ford and Toyota.
One of the major issues, at least right now, is that most of these systems are only reactive, relying on well-marked lanes rather than lines that could be worn or snow-covered. Automakers have already developed technologies that can determine when lane changes are necessary, such as when a collision may be imminent, but they’re also only responding, rather than anticipating something such as an upcoming merge into a turn lane. (Tesla’s Autopilot system, which relies on adaptive cruise control and lane-keeping as other vehicles do, can change lanes but only when the driver taps the turn signal switch.)
For a completely self-driven road system, vehicles need infrastructure that’s under development but still not ready: navigation systems that can detect off-ramps and turn lanes to the centimetre, interactive traffic lights, vehicle-to-vehicle communication systems and “smart” garages to guide cars to parking spots. And that’s even before the lawyers, licensing bureaus and insurance companies figure out who’s responsible for what. We are facing interesting times, indeed.