How does an unmanned car

According to some analysts, in the coming years, a serious transport revolution awaits mankind: cars driven by people driving will be a thing of the past, and unmanned cars will fill the roads instead. Surprised?

There is nothing unusual here. After all, if you take a closer look, the first steps have been taken today: Tesla electric vehicles, for example, use a very efficient computer driver assistant that can keep the car in motion in its lane, at a speed suitable for the current traffic situation, and even able to readjust itself on command.

And all this is ensured thanks to the joint work of several electronic units: ultrasound helps to recognize the presence of other cars on the road, neither rain nor fog are scary to the front radar, and a separate video camera carefully reads traffic signs. The position of the machine in this mode is monitored in real time and is adjusted thanks to GPS.

More advanced automatic driving systems will appear, presumably, in the near future, because already today several large companies are engaged in serious development in this area. You can recall the car from Google, which was created on the basis of Toyota.

By the way, Toyota itself plans in the coming years to launch the production of small two-seater drones that can move on roads at speeds up to 25 miles per hour.

Ilon Musk, an outstanding amateur of innovative developments, does not stand aside, predicting cardinal changes in the coming years (by 2020), thanks to his unmanned vehicles from Tesla Motors.

Even KAMAZ intends to establish production of unmanned trucks by 2025. In fact, it turns out that the massive introduction of unmanned vehicles around the world is only a matter of time.

An example of the successful implementation of unmanned vehicle technology is a Google unmanned vehicle, which for its operation uses, among other things, the Google Street View panoramic street view service.

The basis of the smart system includes: radars, a video camera, a position meter, and a special LIDAR sensor mounted on the roof of the car and quickly scanning the space around it in a radius of 60 m.

The camcorder near the rear view mirror tracks traffic signals and moving objects. Bumper radars prevent any contact of the vehicle with obstacles. The speeds, directions of movements and sizes of objects around, as well as the distance to them - all this information is collected by radars.

As a result, the car is endowed with the ability to promptly and adequately respond to various obstacles in its path. A separate sensor is connected to one of the wheels, it helps to reflect the current position of the car on the map.

GPS coordinates are obtained with high accuracy in the “satellite visibility” mode, and when geostationary satellites are unavailable, standard GPS accuracy is used.

The use of the LIDAR sensor in a Google mobile allows the on-board computer to programmatically recreate a three-dimensional map of the terrain, and, comparing it with the data received from the sensors, weigh the pros and cons when choosing the trajectory and nature of the upcoming movement.

The program will evaluate the current traffic situation, calculate the probabilities of various behaviors of other road users, and, based on all the data at its disposal, will set the drone to the trajectory for the most efficient and safe traffic. Traffic signs and even gestures of traffic police will be recognized and taken into account immediately.

It is noteworthy that when the Google car was just being developed, its first tests were carried out by engineers on a computer model of motion.It was an ordinary road with relatively little transport activity.

Here the google car coped easily, because it was enough for him to turn, turn, accelerate or brake where necessary. But with the transfer of the model to the conditions of a large settlement - everything has changed.

How to respond to an accident ahead? Do I skip a school bus? How to react to the gesture of a traffic police officer, but how about a pedestrian? What makes the appearance of policemen with beacons on among traffic participants? And so on and so forth. Obviously, the flaws come to the surface one by one.

For example, since objects are recognized by comparing the current landscape with previously taken photos, in the changing weather conditions (snow, rain), it is no longer easy for a computer to distinguish a dog from a flower bed or a pillar from a person. Nevertheless, even such an imperfect result achieved cost developers significant effort.

According to the company's engineers, the information that each Google car receives in real time is transferred to a common database, with which all Google cars are connected and can exchange information.

Thus, the google car is able to overcome any difficult situation on the road, except perhaps very unusual ones, in which it remains for it to simply brake. For example, the problem of reaction to birds became a daunting task for Google engineers when they studied the interaction of a car with various obstacles on the road.

In the early stages, the car slowed down every time it “saw” a bird, perceiving it as a serious obstacle. Later, the problem with the birds was resolved, as well as the problem of the pedestrian, who is unclear whether he is going to cross the road, or simply stands about his business near the curb.

Later, a google car was taught to slow down and give signals if a situation is outlined that could even lead to a road accident with a small degree of probability.

One way or another, not far off are the days when unmanned vehicles will supplant cars with drivers. This will happen exactly when unmanned vehicles 100% prove their safety.

Already, there are statistics showing that accidents with Google unmanned vehicles occurred mainly through the fault of another driver, who is a living person. Thus, the prospects for unmanned vehicles are uniquely bright, because the electronics are not affected by emotions, unlike humans.