Google’s self-driving cars are programmed to exceed speed limits by up to 10mph (16km/h), according to the project’s lead software engineer.
Dmitri Dolgov told Reuters that when surrounding vehicles were breaking the speed limit, going more slowly could actually present a danger, and the Google car would accelerate to keep up.
Google’s driverless prototypes have been widely tested on roads in the US.
The UK will allow driverless cars on public roads from 2015.
Google first announced its driverless car division in 2010, and has been testing its technology in modified cars built by other manufacturers.
The cars have travelled on more than 300,000 miles of open road, mostly in California.
In May, the US tech firm said it would start building its own self-driving cars.
The bubble-shaped vehicles will seat two people, propulsion will be electric, and to begin with they will be limited to 25mph (40km/h) to help ensure safety.
In July, the UK government announced that driverless cars will be allowed on public roads from January next year.
In addition, ministers ordered a review of the UK’s road regulations to provide appropriate guidelines.
This will cover the need for self-drive vehicles to comply with safety and traffic laws, and involve changes to the Highway Code, which applies to England, Scotland and Wales.
How do driverless cars work?
The label “driverless vehicle” actually covers a lot of different concepts.
Indeed, the cruise control, automatic braking, anti-lane drift and self-parking functions already built into many vehicles offer a certain degree of autonomy.
But the term is generally used to refer to vehicles that take charge of steering, accelerating, indicating and braking during most if not all of a journey between two points, much in the same way aeroplanes can be set to autopilot.
Unlike the skies, however, the roads are much more crowded, and a range of technologies is being developed to tackle the problem.
One of the leading innovations is Lidar (light detection and ranging), a system that measures how lasers bounce off reflective surfaces to capture information about millions of small points surrounding the vehicle every second. The technology is already used to create the online maps used by Google and Nokia.
Another complementary technique is “computer vision” – the use of software to make sense of 360-degree images captured by cameras attached to the vehicle, which can warn of pedestrians, cyclists, roadworks and other objects that might be in the vehicle’s path.
Autonomous vehicles can also make use of global-positioning system (GPS) location data from satellites, radar, ultrasonic sensors to detect objects close to the car and further sensors to accurately measure the vehicle’s orientation and the rotation of its wheels, to help it understand its exact location.
The debate now is whether to allow cars, like the prototype unveiled by Google in May, to abandon controls including a steering wheel and pedals and rely on the vehicle’s computer.
Or whether, instead, to allow the machine to drive, but insist a passenger be ready to wrest back control at a moment’s notice.