How do Google self-driving cars work?
September 11, 2015 at 11:49 AM
You may not know, but Google has developed many technologies for its self-driving car.
Here is a list of the 10 most amazing technologies.
Refer to schematic and the numbers:
1) Laser Range Finder
Probably the most amazing technology on the car which is the essence of self-driving cars the Lidar (light-radar), with its 64 laser beams creates 3D images of objects and calculates potential hazards and distances between the objects and the point of origin. This system has a range of 200 metres.
2) Front Camera for Near Vision
A camera mounted on the windscreen helps the system notice obstacles and road information such as signals and traffic lights which are interpreted by the car’s navigation software.
3) Bumper Mounted Radar
The car is also equipped with proximity sensors in the front and rear bumpers. In principle it works the same way as a cruise control system would. In essence, the software applies safety distances based on the speed and distance on the vehicle, regulating its speed to avoid any crashes or bumps.
An aerial on the back of the car receives information about the precise location of the car. The car’s GPS’ INU (Inertial Navigation Unit) works in conjunction with the sensors to localise the vehicle. However, since the GPS positioning is not 100% accurate the location is compared with sensor map data previously collected.
5) Ultrasonic Sensors on Rear Wheels
An ultrasonic sensor on one of the rear wheels helps the tracking of the car’s movements and will alert the CPU of any obstacles at the rear of the vehicle. This technology is already used by vehicles that use the automated park assist technology.
6) Devices within the car
The car is fitted with many indicators such as altimeters, gyroscopes and tachymeters to increase the precision of the localisation.
7) Car CPU
The car CPU synergises all the instruments data and gathers them to let the car drive safely by itself.
The advanced software that controls the vehicle has been programmed to observe and interpret road behaviour and other road users signs, it also interprets the types of vehicle present on the road based on parameters such as their size and speed in order for the car to drive safely.
9) Advanced Mapping
Before testing a self-driving car, a regular car was used to map the route and its road conditions and signals. As the car moves the Lidar system starts making records of the environment and compares it to the pre-existing map to figure anomalies such as pedestrians and other motorists in order to engage evasive manoeuvres.
10) Real life road behaviour
Parameters of real life behaviour have been added to the software in order to understand human driving reactions such as 4 way junctions etc.