If technology grows at its current rate, your car will soon do all the driving and concentrating for you. Automakers are currently working on new technologies that will allow cars to drive themselves. They are also modifying existing technologies, such as self-parking and pre-security systems, to make driving safe.
Figure 1: Components to be used in the Driverless Car
Studies show that driver error is the most common cause of traffic accidents. Chris Urmson, who currently works on driverless car technology, goes even further and says that the least reliable part of the car is the driver. Vehicles have become safer, smarter and stronger, but the driver problem still exists.
Car adverts sell us the idea of warm, sunny days and driving through the countryside with a cool wind blowing through our hair, but the reality is quite different. Driving these days mostly consists of sitting in traffic or in the rain, paying more attention to our phones than our surroundings. With the advancement of technology, these problems will not disappear. Almost 2 million people die on the world's roads every year. So if the driver isn't paying attention to the road, who is?
In this article, we'll talk about the technology behind cars that can operate without minimal driver intervention. Also, how far are these cars from mass production and when will we let machines take over?
The primary technology
Google has been working on driverless cars since 2009. Amazingly, these cars have traveled more than half a million miles (804,672 km) without a single accident! While human drivers are involved in accidents every half a million kilometers.
The technology uses a Chauffeur system called LIDAR (light detection and ranging). LIDAR works like radar and sonar, but with more precision. What it actually does is map points in space using 64 rotating laser beams that take more than a million measurements per second and form a 3D model in your computer brain. The system also includes preloaded maps that tell you where stationary things are, like traffic lights, crosswalks, sidewalks, etc., while LIDAR depicts the landscape with moving objects like people and traffic.
Figure 2: Driverless car operation project
To work safely and efficiently, self-driving cars need to understand their position on a GPS map and also the relative position of other cars and pedestrians. While pedestrians or construction increase the level of complexity of the algorithms and information received by cars. Police cars, cyclists and school buses must be handled exclusively by car.
Google collects data that shows the behavior of pedestrians, cyclists and drivers. Around 5 million kilometers of tests are carried out every day in simulators using this technology. As more and more data is collected, situations can be better predicted by the car.
How do you see the road and its surroundings?
The vehicle starts by understanding where it is in the world, taking information from the map and sensor data and aligning the two. Plus, he covers what he sees at the moment. Like other vehicles, pedestrians in the vicinity.
But the driverless car has to do better than just understand its surroundings. You have to be able to predict what will happen.
For example, a truck in front will make a lane change because the road in front of it is closed. The driverless car needs to know this. But in fact, even knowing this is not enough.
What he really needs to know is what everyone on the road is thinking. Plus, the car needs to figure out how to respond in the moment. For example, which trajectory to follow, if you slow down or accelerate at the moment. When combined, all of this becomes quite complicated and is achieved by thousands of algorithm checks.
When this concept began in 2009, it was a very simple system, where the car was driving on the road and had to understand where it is and approximately where the other vehicles on the road are. It was practically like a geometric understanding of the world.
Although this is not what we encounter in our daily lives. On city streets, the problem takes on a whole new level of difficulty. There are pedestrian crossings, cars traveling in all directions, as well as traffic lights and construction on the road.
Above all, you have to respond differently to police vehicles and school buses. Furthermore, the car has to understand when the police officer signals to stop and go.
Chris gives a wonderful example where the car is stopped at a red light. An ordinary driver sitting in the car cannot see the cyclist on the far left of his vision. But the driverless car can! This is possible due to the laser data scanned by the car from the nearby area.
Now the cyclist is passing along the road. His signal turned yellow, but he keeps moving. Now, halfway through, his light turns red and ours turns green. Most drivers move on because they didn't notice the cyclist. But the driverless car, because of its laser data, anticipates that the cyclist is passing by. He responds confidently to this. As the other drivers begin to advance, the cyclist narrowly manages to escape, avoiding the collision, but the driverless car waits patiently for the cyclist to cross.
Limitations
Although people working on this technology are confident that it will eventually reach the market, it still has some serious limitations.
To date, available sensors and artificial intelligence are not capable of seeing and understanding the vehicle's surroundings with the same precision that a human being can.
For example, if a ball rolls down the road, a human might anticipate that a child might follow it. Artificial intelligence cannot provide this level of thinking nor can it communicate with its surroundings in real time. Reaching this level of artificial intelligence would take almost 16 years.
As of 2014, the latest prototype had not yet been tested under heavy rain and snowfall. This was mainly due to safety concerns, as cars are pre-programmed with route data and do not obey temporary traffic lights. The vehicle faces difficulty in understanding when a trash can or trash on the road is harmless.
LIDAR technology cannot detect some potholes or distinguish when a human, such as a police officer, is signaling the car to stop. Google expects these issues to be fixed by 2020.