This is how Google’s driverless car prototype actually works
Next year, driverless cars will be tested in three cities in the UK, but how do autonomous cars work?
Google has been testing its prototype cars on US roads-not yet trialled in the UK-and revealed some details about how its self-driving cars work.
Here, we explain some techniques.
Driverless cars are already here …
Much of the autonomous technology used in Google’s autonomous cars is already on the road.
You may have seen ads advertising the Volkswagen Polo automatic braking or Ford Focus automatic parallel parking, which are using the increasing use of proximity sensors for parking.
Combine these sensors with the automatic steering technology used for parking, put on the seemingly old-fashioned technology, which is cruise control, and you will get loose frames for a self-propelled car.
How many sensors does the car have, and what do they do?
Google’s driverless cars have eight sensors.
The most striking is the rotating rooftop LiDAR, which is a camera that uses 32 or 64 laser arrays to measure the distance to an object to create a 3D map within a range of 200m to let the car “see” the danger.
The car also has another set of “eyes”, which is a standard camera that can be aimed at the windshield. This will also find nearby hazards, such as pedestrians, cyclists and other motorists, and read road signs and detect traffic lights.
For other drivers, the radar installed on the bumper has been used in the intelligent cruise control system, which can track the vehicles in front of and behind the car.
Externally, the car has a rear antenna that receives geographic location information from GPS satellites, and an ultrasonic sensor on one of the rear wheels for monitoring the car’s movement.
There is an altimeter, gyroscope and tachometer (tachometer) inside the car, which can make more precise measurements of the car’s position. Together, these provide the car with the high-precision data needed for safe operation.
How Google’s driverless car works
No individual sensor is responsible for making Google’s autonomous car work. GPS data, for example, is not accurate enough to keep the car on the road, let alone in the correct lane. Instead, the driverless car uses data from all eight sensors, interpreted by Google’s software, to keep it safe and take it from A to B.
Data received by Google software is used to accurately identify other road users and their behavior patterns, in addition to commonly used road signs.
For example, Google’s car can successfully identify a bicycle and understand that if the rider extends an arm, he intends to do a maneuver. So the car knows how to slow down and give the bike enough room to operate safely.
How Google Autonomous Cars Are Tested
Google’s autonomous vehicles, of which it has at least ten, are currently being tested on private tracks and, since 2010, on public roads.
The car always has two people inside: a qualified driver with impeccable registration sits in the driver’s seat, to take control of the car by turning the steering wheel or pressing the brake, while a Google engineer sits in the passenger seat to control behavior. software
Four US states USA They passed laws that allow the use of driverless cars on the road, and Google has made the most of it, testing its car on freeways and suburban streets.
Steve Mahan, a California resident who is blind, participated in an exhibition driving test, in which the car drove him from his home through the city, including a visit to a self-service restaurant.
However, it is not about telling your car where you want to go, sit back and relax.
Google software engineer Sebastian Thrun explained in a blog: “Any test begins with the dispatch of drivers driving traditional cars to draw routes and road conditions.” By mapping such as lane signs and The function of traffic signs, the software in the car can be familiar with the environment and its characteristics in advance. “
Are driverless cars safe?
This is one of the questions that still arises in the debate about a car without a driver: is it safe to give the robot control of the vehicle?
Proponents of the technology of self-propelled cars quickly show statistics that show how dangerous roads are at the hands of non-autonomous cars – in 2013, 1,730 people were killed in Great Britain as a result of car accidents, and another According to the Office of National Statistics, 185,540 people were injured.
Numbers around the world are just as frightening, and fatal accidents have killed 1.2 million people last year. Google claims that over 90% of these deaths were the result of human error.
In April, Google announced that cars without a driver drove over 700,000 miles (1.12 million kilometers) without a registered accident caused by one of his vehicles – one was hit from behind, but the other driver was guilty.
Although this is an unbelievably small number compared to how many British drivers miles per year – in 2010, Admiral insurance company suggested the number could be close to 267 billion miles – the fact that Google autonomous cars are still free from accidents is encouraging.