How does a self-driving car work?

By T.S.

CARS that can drive themselves, a staple of science-fiction, have started to appear on roads in real life. Google’s self-driving vehicles are the best-known, but most carmakers are also developing them. In 2011 BMW sent a robotic car at motorway speeds from Munich, the German carmaker’s hometown, to Nuremberg, about 170km away (with a driver on board just in case). Audi got a self-driving TTS Coupe to negotiate 156 tight curves along nearly 20km of paved and dirt road on Colorado’s Pikes Peak, with nobody behind the wheel. Proponents say that driverless cars would reduce road deaths, ease congestion, reduce fuel consumption, improve the mobility of old and disabled people and free up time spent commuting. So how do they work?

In many ways self-driving cars are a logical extension of existing driver aids such as lane-keeping systems (which follow road markings and sound a warning and correct the steering if a vehicle starts to drift out of its lane), adaptive cruise control (which maintains a constant distance from the vehicle in front, rather than a constant speed), auto-parking systems (which can reverse a car into a parking space), emergency braking (which slams on the brakes if an obstacle, another vehicle or a pedestrian is detected in front of the car) and satellite-navigation systems. Computerised control of a car’s steering, acceleration and braking is already possible under some circumstances, in other words. For a car to drive itself, these systems must all be tied together using software, and supplemented with a set of sensors so that the software can tell what is going on around the vehicle.

Accordingly, today’s self-driving cars are covered with sensors. Mapping nearby features, spotting road edges and lane markings, reading signs and traffic lights and identifying pedestrians is done using a combination of cameras, radar and lidar (which works like radar, but with pulses of light rather than radio waves). Ultrasonic detectors provide more accurate mapping of the surroundings at short range, for example when parking. Gyroscopes, accelerometers and altimeters provide more accurate positioning than is possible using global-positioning system (GPS) satellites alone. Google’s cars scan their surroundings to build a detailed 3D map of features such as road edges, signs, guard-rails and overpasses. Each time a car follows a particular route, it collects more data to update the 3D map. Google’s software also ingests data on speed limits and recorded accidents. Because the car’s roof-mounted sensors can see in all directions, it arguably has greater situational awareness than a human driver. Google’s self-driving cars have clocked up 700,000km (435,000 miles) under autonomous control without incident. Snow-covered roads and temporary signs around roadworks still pose challenges, but the technology is improving all the time.

The BMW i3, an electric car capable of driving itself in stop-go traffic (by linking adaptive cruise control with lane-keeping) is due to go on sale later this year, but most observers expect it to be a few years before entirely autonomous vehicles are available. The transition is likely to be gradual. There is a continuum between a fully manual and a fully autonomous vehicle, and over the coming years it will slowly become possible to outsource more and more driving tasks to your car, particularly as tightening safety standards make driver aids compulsory. If fully autonomous vehicles do eventually become widespread, there could be dramatic consequences for car design, car ownership and urban planning. Some places might even ban manual driving, to save lives and ease congestion. Self-driving cars will not arrive overnight, but they are on their way.

(Picture credit: Google)