Transport accidents has become one of the destabilizing factors of modern cities and towns
Innovations in the field of computer simulations can help reduce the destabilizing effects of accidents and chaos caused by inappropriate driving
This primarily refers to the modern teaching methods and achievements of driving experience with special software
There are many examples of optimal solutions to this problem in Israel
... First grade Smart Drive appeared several years ago in Ashdod. After the successful launch of a branch school has opened in the Azrieli mall - in Tel Aviv, the city where he learned to drive, probably the most difficult in Israel. Within a few years of its existence as a driving simulator dubbed the domestic realities. With the main emphasis on multitasking. Flick of the wrist virtual machine can become a light truck, a private car with a manual transmission, or the traditional Israeli machine. The main plus - driving simulator in as close as possible to reality.

Said Chairman of the Board of Directors of the "Smart Drive", "The Israeli roads are much more dangerous than we can imagine. Programs in our simulations incorporated a variety of situations that drivers face every day, but to be faced during a standard driving lesson is almost impossible. This is very important for a person who does not want to just get the rights, but to learn to drive. "

Besides the normal passenger cars can try to manage a bike. For this simulator, a special program with all aspects of driving, "tustusa" in the city and intercity routes. In this simulator actually allows you to feel the road.

Says CEO "Smart Drive": "We teach to what the car can be difficult to learn. For example, if a person begins to learn in the summer, the experience of driving in winter conditions (and this rain, sleet in the north, and much more), he simply will not, or vice versa. simulator can make a mistake, go back a few steps back, analyze the situation and more a mistake to avoid. Error on the road is often the last. A simulator allows to reduce the number of errors to a minimum.

One of the biggest advantages of the project is that, at least theoretically, because these machines can reduce the number of lessons on these machines. Especially considering the fact that the standard lesson lasts about 40 minutes, and a lesson on the simulator for an hour.

Said driving school student: "Feeling great. Before I drove, but it was a long time, the simulator is very similar to a real machine. There is a sense of the road, traffic, everything in fact. I loved it."

The main thing that needs to bring to the realities of the Israeli use of simulators - this is the very culture of driving, which we all have been talking a lot. After all, much easier to understand the safe driving pretty work-out in the classroom on the machine on which the accident may even be an element of training. But already on this car is not even a thought scorcher or break the rules. For it is clear than it might.
In the U.S., this solution has a much bigger scale and presented a more promising solution
Motor Vehicle of the U.S. state of Nevada on May 7 approved the issuance of Google's first-ever license to test the autopilot controlled cars on public roads. Opportunities unmanned Toyota Prius, equipped with an arsenal of different sensors for Google successfully demonstrated on state highways, residential areas, Carson City and downtown Las Vegas.
According to a press release from the Office of Vehicle Nevada, the first test cars with autopilot on Google will receive the red license plates. The left side of the plate will be located infinity symbol - a symbol of "the car of the future", said the head of the Office of Bruce Breslow (Bruce Breslow).
Original red numbers will be installed only on the experimental unmanned vehicles, making them easy to recognize as ordinary citizens, and representatives of law and order, "- said Breslow. For the first series of unmanned vehicles license plates will make the green with the same characteristic symbol of infinity.
It should be noted that Google was not the first company that decided to invest in the development of unmanned vehicles. Now these studies are carried out General Motors, Ford, Mercedes-Benz, Volkswagen, Audi, BMW and Volvo. Company Cadillac, in particular, is working on a semi-automatic control system Super Cruise, which is scheduled to be released on the market in the next few years.
The first major test of unmanned vehicles began in the 1980s in Germany. Then the military experts of the University of the Bundeswehr Munich tested vehicle Mercedes-Benz, equipped with a robotic control system based on the video sensor. Vehicle could accelerate to 100 kilometers per hour, but the test were in the empty streets - in fact, far from the reality conditions.
"Guglomobil" is equipped with a fully autonomous control - robotic system is responsible for turning steering, acceleration and braking. Safe movement on roads in the real world (with traffic lights, intersections, cars and pedestrians) is provided by a satellite navigation system, the service Google Street View and a whole range of sensors. These include a rooftop optical sensor LIDAR, determining the distance to objects by reflected light, radar on the front of the car, the sensor positioning on the rear wheel, and of course the camera in the car.
Lead engineer responsible for the development of unmanned vehicle Google, is vice-president Sebastian Thrun (Sebastian Thrun), who heads the Laboratory for the Study of Artificial Intelligence at Stanford University. In 2005, a team of scientists led by Troon won the competition of unmanned vehicles, organized by the Office of Advanced Research Projects (DARPA) of the Ministry of Defense. In October 2010, Troon announced the development of an unmanned vehicle on behalf of Google.
The first test "guglomobiley" began in California. Engineering samples of the successfully passed the Google campus in Mountain View to the office in Santa Monica, and then to Hollywood Boulevard. Subsequently drones swept over the bridge, "Golden Gate" in San Francisco, the highway SR 1 running along the coast of the Pacific Ocean, and on the shores of Lake Tahoe.
At present, according to the Las Vegas Sun, the park "guglomobiley" includes at least eight cars - six "hybrid" Toyota Prius, one Audi TT and a Lexus RX450h. By the end of March 2012, a total of "guglomobili" overcame more than 320,000 kilometers. During testing one such unmanned Prius even brought a blind man in a diner Taco Bell.
In the entire history of tests "guglomobil" got only one incident - unmanned Prius hurt ordinary "manned" Prius in Mountain View. Representatives of Google, however, immediately declared that "guglomobil" at the time of the accident was in manual mode and automatic control system working smoothly.
Be that as it may, Google has become the first in the United States (and the world) company, has officially admitted to the trial operation of its unmanned vehicles on public roads. It has become possible after the approval in February 2012, legislation authorizing the movement of autonomous vehicles on the roads of Nevada.
Under the law, a license to travel on public roads may only be granted to applicants whose unmanned vehicles dashed when tested at least 16,000 miles. In addition, the movement of such vehicles on public roads during the test in the cabin must be at least two people - one in the driver's seat, and the other - the "navigational" place.
In addition to Nevada such regulations are preparing to review in California, Florida, Oklahoma and Hawaii. Outside the U.S., the fate of unmanned vehicles is not so clear. The Daily Telegraph, in particular, wrote referring to the head of the Parliamentary Committee on Transport UK Robert Gifford (Robert Gifford), that in the kingdom to pass to control unmanned vehicles is not planned.
"Right now, there are cars that can self-park [for example, Volkswagen Touran - ca." Heathcliff "], but only to help the driver, instead of replacing it. Government needs to understand how [unmanned] technology will develop in the next ten years. .. "- said Gifford.
If you believe Google representative Anthony Lewandowski (Anthony Levandowski), quoted by The Wall Street Journal, the company hopes to finalize the "guglomobilya" "faster" than a decade.
Scientists at Columbia University have found that the use of autopilots for cars will help increase the efficiency of transport and passenger transport by at least 273 percent. Study of its kind - is not new, but the remarkable work of the Americans doing what the calculations are based solely on the existing technology.
Imperfect drivers

The first thing to understand and accept - a man, any man, a very bad driver. His main drawback - a very long delay in responding to the changing situation on the road. This is, of course, much to the physiological limitations. For example, a person needs time to analyze the situation and respond to it.
In total, depending on the physical condition of the driver, visibility, concentration and other factors, the reaction time may vary from about 0.15 seconds to tens of seconds. With the average value of a conditional 1.5 seconds and a vehicle speed of 90 kilometers per hour is the response time provides a stopping distance of more than forty yards. But this may be added the time it takes to perform direct command vehicle - such as a full stop after you press the brakes (this adds to the stopping distance is 40 meters).
Because of this man is extremely inefficient use of the roadway. As the analysis of scientists from Columbia University, in a straight highway vehicle covers an average of about 5 percent of the roadway. All of the remaining 95 percent, which, however, is not enough, remain on the imperfect human reflexes.
In addition, according to scientists, it is necessary to take into account the psychological characteristics of drivers, who are also not in favor of playing the quality of driving. For example, a motorist is quite able to lock in the heat of the moment decision by one or two bands, which in some cases can lead to a complete stop vehicles on the highway.
Perfect movement

At present, work on the improvement of road maintained, relatively speaking, in two ways. The first (which we call the global) is trying to improve the traffic outside, without considering the peculiarities of human drivers. In these models, the car is passive, in fact, he plays the same role as the fluid simulation in the water current. This direction can be safely called a classic and fairly well developed - the scientists used computer simulation, for example, currents with different properties on graphs (represented by the diagram of roads).
Arising in such problems are very complex dynamical systems, in particular, they can abruptly shift from one steady state (the free movement of vehicles) to another (a traffic jam, when the speed of individual cars is close to zero.) Moreover, studies show that the conditions for such a transition depends on many factors, including, for example, the structure of the network of roads.
However, in this area there are some success stories - so widely used in Europe (pdf) different modes of traffic lights, which, in practice, can significantly relieve either way, or stop the growth of the load. In the basis of these methods is the idea that the traffic jams on the way to get rid of completely will never be.
Implicitly, most models of this kind mean that drivers in motion are subject to the rules of the road, that is responsive to the signals provided externally, and it is, as mentioned above, due to the psychological characteristics are not always true. Thus, we come to the second class of models, in which the driver is not at all - instead, set a fundamentally different task organization of the movement, in which cars are considered as the active part of the model.
On the one hand, it makes it easier in theory, but on the other - complicates its practical implementation, because the interaction between a car or with some kind of control system is necessary to organize.
Interaction as a way to success

In 2011, the Korean mathematics with interesting results. It turned out that if the cars will simply inform each other in the direction of movement and speed (this is, of course, implies some inter-vehicle communication - signal lights there will not be enough), then you can avoid traffic jams. Local interaction is often enough to facilitate the transport situation.
In the new work, which appeared in the magazine IEEE Spectrum, scientists considered more radical scenario. They assumed that the cars are controlled autopilot. Here we consider the simplest case - straight highway with no intersections and junctions. It turned out that in this case, one lane of the road may miss up to 12,000 vehicles per hour at a speed of 120 kilometers per hour.
For comparison, if the car is driven people, capacity, other things being equal, would amount to 2-3 thousand avtomobiley.Vse case, according to the researchers, is that at this rate the drivers usually keep a distance of 40-50 meters (this is considered safe). In addition, researchers found that, if the machines are not full autopilot, and systems such as adaptive cruise control, a safe distance at a speed of 120 kilometers per hour can be reduced to 25-30 meters. In this capacity grows "only" a factor of 1.5-2 to 4 thousand vehicles per hour.
However, the main advantage of the new work is that scientists used to analyze existing technologies. Detail the development of these technologies "Lenta.ru" already written.
There is already a
In June 2011, Volkswagen has revealed the development of "Temporary Autopilot" (Temporary Auto Pilot, TAP). In essence this is the equivalent of the cruise control, which can provide the bandwidth 4000 machines, which was discussed above. The system consists of an onboard computer and a large number of sensors: radar, cameras, laser scanners and other devices. Such temporary autopilot can hold the car on the road in traffic at speeds up to 130 kilometers per hour. In this he follows the layout, understand some speed limit signs (according to them, he chooses the speed limit).
In August of the same year his autopilot presented by BMW. To the functional development of its system is superior Volkswagen (at virtually the same technical component), because it allows the machine yourself to accelerate, slow down, overtake, go to the adjacent roads and wedged into the flow going, and adapt to road conditions. At the time of the official announcement of the test car with the autopilot was 4800 miles.
In January 2012, Volvo has shown the public system, which allows the machine yourself to join in the column. From the point of view of modeling, unlike previous systems, the development of Volvo cars involves interaction with each other in the sense that one vehicle understand the behavior of another.
In April 2012, the company showed its development Cadillac. Its system to the functional and technical component is similar to the development of BMW. Car with such a system can also accelerate and brake. In contrast, however, from previous developments Cadillac calls the date from which plans to install such systems on their cars - in 2015.
Finally in May 2012 held a truly revolutionary event - control vehicles of the U.S. state of Nevada allowed the company to experience Google autopilot controlled cars on public roads. In addition to Nevada such regulations are preparing to review in California, Florida, Oklahoma and Hawaii.
In conclusion

No need to explain that the autopilot enormous potential.