Lecture
Cars have existed for a long time and during all this time only design power and efficiency changed in them, but in general everything remained the same, but soon a big evolutionary step forward awaits them. A few years ago, experts considered completely unmanned vehicles a matter of a very distant future, but then, suddenly, a revolution occurred in the development of neural networks, that is, algorithms that can learn to become better and this gives us the opportunity to create completely unmanned vehicles in the near future. Because today, many companies are already developing their unmanned products for the mass market including Tesla, Uber, General motors Volkswagen, Audi, BMW, Volvo, Nissan, Google and others.
For unmanned vehicles in total there are five levels of automation, respectively, the first is the most primitive and the fifth is the most advanced. Today we are almost at the fourth level.
The Automation Automation Classification was developed by the Society of Automotive Engineers (SAE) and contains 6 levels:
Level 0. No automation, the driver does all the work.
Level 1, “hands on”, “driver assistance”. The driver and system drive the car together. Example: the driver drives, and the system adjusts the engine power while maintaining the set speed (cruise control) or adjusts the engine power and controls the brake, while maintaining the set speed and, if necessary, reduces it to maintain the distance (adaptive cruise control). Another example is automatic parking (en: Automatic Parking), when the speed is determined by the driver and taxiing is automatic.
Level 2, “hands off”, “partial automation”. The system fully controls the car, carrying out acceleration, braking and taxiing. The driver monitors the ride and is ready to intervene at any time if the system cannot respond correctly. Despite the name “hands off,” such systems often require the driver to keep their hands on the steering wheel, as a confirmation of their readiness to intervene.
Level 3, “eyes off”, “conditional automation”. No immediate response is required from the driver. He can, for example, write messages or watch a movie. The system itself responds to situations that require immediate action, such as emergency braking. The driver is required to be ready to intervene for some limited time determined by the manufacturer.
Level 4, “mind off”, “wide automation”. It differs from level 3 in that it does not require constant attention from the driver. For example, he may go to bed or leave the driver's seat. Fully automatic driving is carried out only in some spatial areas (geofences) or in some situations, for example, in traffic jams. Outside of such places or situations, the system is able to stop driving and park the car if the driver has not taken control.
Level 5, “steering wheel optional”, “full automation”. No human intervention is required.
This level allows the driver not to keep constant attention on the road and gives the person the opportunity to be distracted by their affairs.
For example, Yandex recently demonstrated its unmanned vehicle the level of automation of which was close to the fourth and this during the demonstration already allowed the driver not to sit in the driver's seat. Tesla cars that drove everything from the US from the west coast to the east had a second level of automation, which requires constant concentration on the road from the driver. But Elon Musk, the head of Tesla, expects that he will have fully unmanned fifth-level vehicles next year.
According to him, the widespread introduction of robomobiles will provide tremendous benefits and benefits for the population. Firstly, you can earn good money on your car. Today, on average, a car is used for 10-12 hours a week, most car owners spend one to two hours a day driving, and the rest of the time the car is idle and does not bring any benefit. But if your car is self-driving then it can no longer stand idle in the parking lot. While you are sitting in the office, you can send your car at a charge. Make the car work 50-60 hours a week so it will increase the car’s efficiency by 5-6 times. but the most impressive thing is that Tesla plans to make the cost of traveling one kilometer on Robotaxi more than 10 times cheaper than today at Uber. such a reduction in price will occur, among other things, due to the fact that today 80% of taxi expenses are driver expenses for his time spent, because companies dream of removing this expense.
Part of the released money can be spent on lowering prices for end consumers, thus equipping the car with an unmanned control system will fully pay off due to increased demand for trips. which will happen not only due to lower prices but also because of increased convenience and reduced social stress. due to the absence of a stranger in the car in the form of a driver. who may be talkative or tired of having a bad mood driving too fast and so on. in any case, such is the plan of Elon Mask — he wants to do so in order to buy some other car than unmanned Tesla for a person was madness. Since, according to him, unlike simple unmanned gasoline vehicles, it’s precisely unmanned electric vehicles that are the most efficient on the roads at the cost of driving and generating income from their car. Residents of the United States only with the purchase of an electric unmanned car will be able to fully pay for it in less than two years, while doing nothing. In other countries, of course, it will take more time, due to lower prices for taxi services. according to calculations, the mask per year on its unmanned if in the US it will be possible to earn 30 thousand dollars.
The company Elon Mask will take 25-30 percent of the revenue. Customers will order your car through the application on the phone. Through the same application, you can connect or disconnect your car from the taxi fleet in one click. Thus, it will be possible to get a little more than 300 thousand dollars from your car in 11 years. Elon Musk believes that any non-electric taxi robe will be uncompetitive. Of course, this alignment could lead to a new gold rush like the recent crypto rush. which can simply disastrously affect the cost of taxi services because if all people can send their cars in taxi mode. If the cost of a kilometer traveled is really 10 times cheaper than with ordinary taxis today, people will start buying entire fleets of unmanned cars to make more money, and as a result, the market will begin to oversaturated due to which powerful dumping of prices can begin, as a result, moving around taxi will be cheaper than today by public transport. By the way, therefore, some experts think that public transport has no future. And of course, as always with the gold rush, sellers of shovels, that is, automated cars, will earn the most. Of course, all this sounds like a fantasy, but many people really believe in such an outcome. however, not everything is so bad, because such a turn will have a very positive effect on the approximately demand for deliveries of goods from stores. If you have the opportunity to buy products online and the store will send them to your home on a robomobile. This is due to the extremely low cost of delivery that you will use these services more and more often so often that in the end you will almost never physically walk around the store again. Even if you need to buy just a couple of loaves of bread.
Startup Nuro
This can lead to a whole market for unusual services. For example, an unmanned cinema, an unmanned beauty salon, this seems ridiculous, but the concept of an unmanned store has already been introduced by some companies.
Robomart is the world's first unmanned electric vehicle-based grocery store with wireless charging [CES 2018]
Today, several startups are engaged in the creation of self-governing stores without security cashiers and generally no one inside. you can call the store through the application, it arrives at your home, recognizes your face, opens its doors and you buy whatever you like. it also opens up new prospects for marketing.
Suppose you recently searched in the Google search engine for the name of some shoes that you would like to purchase. But they didn’t do this, postponing the next time. However, the search engine remembered your request and suddenly, after a few hours, you get a little stupefied on the phone saying that the shoes you wanted to buy right now are in an unmanned store, which, supposedly by accident, is standing near your house. Now the chance that you immediately buy these sneakers increases significantly and there is almost no doubt that marketers will use such tricks. Let's consider how robomobiles can still benefit us.
Commonly installed sensors:
Unmanned vehicle software may include machine vision and neural networks
Some systems rely on infrastructure systems (for example, built-in on or near the road), but more advanced technologies allow you to simulate the presence of a person at the decision-making level about changing the position of the steering wheel and speed, thanks to a set of cameras, sensors, radars and satellite navigation systems.
Some experts believe that unmanned vehicles optimize traffic flow, which will reduce the number of cars on the road. After all, now the car I have to stand idle until you use them. Thus, to meet the transport needs of the entire urban population, much fewer cars will be needed than is required today, which means traffic jams will disappear. There is a theory that is directly opposite to it, saying that the number of traffic jams and a fleet of cars, on the contrary, can increase significantly because if the cost of unmanned taxi services is much lower than manned ones, then more people will prefer to travel by taxi than by public transport. Because of this, traffic jams will become more. maybe the truth is somewhere in the middle and we generally will not see any changes in the amount of traffic on the roads. there is also an assumption that personal transport will begin to rapidly disappear. Among us there will be very few people who personally will own the car because most will prefer to constantly use a taxi, given its penny cost and the lack of the need to tinker with maintenance papers. And this, today, is one of the global trends in the development of a shared economy, the so-called sharing-economy. when the population refuses to acquire property benefits in order not to bear responsibility and costs but continues to have access to all the achievements of scientific progress using their joint consumption. There will be a reduction in the cost of transporting goods between cities, not only due to savings on truckers' wages due to elimination of the time required for drivers to relax, but also due to fuel savings due to the optimization of traffic flow.
Given all of the above, we can say with a certain degree of certainty that almost all the cars on our roads will become unmanned a maximum of 10-15 years.
Of course it sounds just bold for the average person, but only because people don’t think in exponential terms. We are used to thinking linearly and very in vain, since many things in our world are changing with an exponential rate.
These two photographs show the same New York city in 1900 and 1913.
In 1900, only one car moves along its street, and everything else is a horse cart.
But in just 10-13 years, the picture is radically changing. Here it is already extremely difficult to find a horse-drawn cart; almost all of them were replaced by cars. Once again, about 10 years have passed, the industry has completely changed, and as much as one could not believe mechanical carriages says that a horse is forever cars are just temporary pampering. The difference in utility between an unmanned vehicle and ordinary cars is no less significant than the difference between a horse and a conventional car. And therefore the future of robomobiles is predetermined - your child will never receive rights in 15 years, it will be atavism.
Let's consider what other economic and social advantages besides those that were listed earlier, unmanned vehicles will give us. Firstly, the cardinal minimization of road accidents and the almost complete exclusion of human casualties, at least among passengers inside the car. Hence the significant reduction in the cost of insurance, quick-response medicine. Due to this, only in the USA alone it is estimated that more than 190-300 billion dollars will be saved annually.
The European Parliament proposes to consider the special legal status of the electronic personality for complex robots that make independent decisions - then they can be blamed for compensation for the damage caused by them. In fact, by this decision, greedy European deputies will remove responsibility from manufacturers of software for artificial intelligence systems used in autopilot cars. The most urgent need for regulation is the sector of unmanned vehicles, according to a resolution of the European Parliament. One of the proposals of the European deputies is to create a special fund, by analogy with car insurance, from which damage would be compensated when there is no usual insurance coverage. The European Parliament called on insurance companies to develop new products that are consistent with the development of robotics. Auto insurance covers only human actions and errors, and the insurance system for robots should take into account all possible liability.
The most interesting thing in the development of auto-driven cars is that there is still not a single train or airliner (airplane) without pilots, on the contrary there are 2-3 drivers (pilots) on each flight in order to minimize any problems. Thus, corporations are more important than money, rather than specific human lives. And you would give your life, trusting an unmanned taxi, in which Uber saved on a driver’s fee?
In modern unmanned vehicles, algorithms based on the Bayesian method of simultaneous localization and mapping are used (SLAM, simultaneous localization and mapping). The essence of the algorithms is to combine data from car sensors (real-time) and card data (offline). SLAM and the method of detection and tracking of moving objects (DATMO, detection and tracking of moving objects) are developed and used in vehicles of the Google Waymo subsidiary. Google has sued Uber over theft of the latest technology from Google. However, since 2017, Google has put the SLAM library openly for free use by any third-party company.
As stated above, AI is the heart of the car. AI detects objects from cameras, tries to guess who it is (dog, person, car, traffic sign, etc.), how pedestrians and other cars will behave. For such artificial intelligence to work, engineers “feed” him huge amounts of data so that special algorithms can be trained on this data. The more quality input data, the better the algorithms will work.
Although the algorithms have advanced far, they are still stupid like a 2 year old child. A striking example is the incident with the Uber UAV (due to which a person died), the algorithm could not recognize the person on the road (in other things, as the driver did not have time to notice him). But besides a person, one must also “see” many other objects - every car, road sign, traffic light, be able to determine lanes and many other things.
Unmanned vehicles require new road infrastructure. And not just infrastructure, but smart infrastructure in which cars could communicate not only with the infrastructure itself (signs, traffic lights, etc.), but also with other cars. Here are some basic terms:
For example, a car is driving along a highway, and a road sign 300m in front itself says “I am such and such a sign, I am there.” An unmanned vehicle will be able to understand in advance what is ahead and plan its actions in accordance with this information.
Economic benefits
dramatic minimization of accidents and the almost complete exclusion of human casualties (at least among passengers inside the car), hence a significant reduction in the cost of insurance and quick-response medicine;
Social benefits;
Other benefits
disadvantages
So in 2015, a Jeep Cherokee hacked a jeep over the Internet, a hacker took control of a car, after which the manufacturer Fiat Chrysler Automobiles was forced to patch software for 1.4 million devices.
Every year, 1 million 300 thousand people die in traffic accidents on the planet. Any transport manned by humans is a means of increased danger. It is especially frightening when bus drivers fall asleep at the wheel. people is always an element of instability according to statistics in 94 percent of cases of all accidents the human factor is to blame. thus, if we completely forbid them to drive people at the legislative level, the number of accidents during the implementation of the fifth level of automation will immediately decrease by about 90 percent. and a further decrease can be made gradually improving the technical part of the robomobiles. a person cannot simultaneously look in all directions from his car, but unmanned vehicles can. unlike the driver who sits on the driver's seat, the cameras are located on all sides of the car, which allows her to look at 360 degrees. never distracted by anything; the car will never fall asleep while riding in it; there will never be a bad mood, stress, she will never think about any personal problems, reducing her concentration on the road, and her reaction speed is always much higher than any of us. today, relatively primitive autopilot systems, judging by the statistics of kilometers traveled, are about twice as safe as humans.
When introducing the fifth level of automation, they want to raise this indicator several times more. and when the entire fleet of cars will be unmanned and get behind the wheel will be banned at the legislative level. this will allow machines through the network to effectively coordinate their actions reducing the number of accidents to almost zero values compared to today.
Saving the time now spent on driving will allow you to do more important things. for example, start working on a computer while you are in a car, or relax or watch a movie. Also of the obvious advantages of unmanned transport is the transport of goods in hazardous areas during natural and man-made disasters or military operations. But of course, everything has its advantages and disadvantages. With the total introduction of unmanned vehicles there is one significant minus. This deprives a huge number of jobs in the population. Ordinary taxi drivers with their cars will not be able to compete in price with safe robomobiles and thus they will lose their job. In the United States alone, over 2 million people lose their jobs. Truckers will be particularly affected there, as today truck drivers receive about $ 70,000 a year. And this is one of the highest paid jobs among the professions that do not require higher education. thus, average skilled people will lose the opportunity to earn decent money. Chinese startup To Simple recently launched test flights of several unmanned trucks between the US cities of Tucson and Phoenix.
The United States Truckers Union, of course, did not immediately like the idea. Truckers demanded to ban unmanned vehicle tests for fear of losing hundreds of thousands of jobs. in Russia, unmanned vehicles will replace the work of 700 thousand people. and this is only among ordinary taxi drivers who are not truckers, and what does the number of people doing so in the savannah and for me officially trucker in our country have not yet taken into account. unmanned vehicles will not please either, for example, in 2017 in Russia it was registered to hear 925 thousand heavy trucks. if they all become self-governing, then you yourself understand.
Also, with unmanned vehicles there are several moral problems, namely, whose life should the car programmers put above the owner of the car or the pedestrian. Let's say you are driving in your robomobiles on a mountain road and suddenly two people run out onto this very road - it is too late to brake the car is driving too fast. What should your unmanned car do? Crush these two and drive on? Or turn to the side, but fall into the cliff and thereby the killers of the person sitting inside? the answer is obvious from a business point of view - in real life no one will buy a car, knowing that in such a situation, it will kill its owner. That is, robotic cars will be ruthless killers of pedestrians. According to a number of experts, accidents should be borne by car manufacturers if the accident occurs due to a technical malfunction, in which case manufacturers will have an incentive to invest in eliminating such problems not only to protect the image, but also to avoid financial and legal consequences, Elon Musk agrees with this statement. In order to make unmanned vehicles safer than cars with drivers, it is necessary to test them. It can be assumed that not all tests will pass successfully.
However, this does not mean that you need to close your eyes to these risks. On the contrary, autonomous car manufacturers must be held accountable for their mistakes and failures, the publication notes.
In light of the two fatal resonance events in 2018 involving Uber and Tesla Motors unmanned vehicles, MIIT Vice President Xin Guobin said road testing of unmanned vehicles is a complex process in which safety should be a top priority. reports Reuters.
In order to ensure the safety of road trials, we will require that testing take place on separate streets and that throughout the test the driver is behind the wheel and monitors the car and the surrounding environment, ready at any moment to take control of the machine. This is a lesson we learned from accidents with Uber and Tesla, ”he said.
November 8, 2017 in Las Vegas, the solemn launch of the Arma unmanned bus, developed by the French startup Navya, took place. After only a couple of hours from the start of operation, the electric shuttle got into an accident. The accident was provoked by a truck driver: backing up, he crashed into the front wing of an unmanned bus.
The development of unmanned vehicles is accompanied by a number of ethical issues, including: moral, financial and criminal liability for accidents, decisions made by the car before a potentially fatal collision, data protection problems and problems of job loss.
There are a number of opinions regarding who should be held liable in the event of an accident, in particular in the event of injuries. According to some experts, car manufacturers should be held liable in the event that an accident occurs due to a technical malfunction. In this case, manufacturers will have an incentive to invest in the elimination of such problems, not only to protect the image but also to avoid financial and legal consequences. At the same time, there is a contradictory point of view according to which users or owners of unmanned vehicles should be held responsible because they are aware of the risks associated with their use.
The next problem is the question of how self-driving cars should be programmed to act in emergency situations where either passengers or other road users are in danger. A classic example of the moral dilemma facing car manufacturers and software developers is the problem of the trolley, in which the trolley conductor has the choice of leaving it on the original route and knocking down 5 people or moving the trolley to a siding and knocking down one person. In this problem, two main questions must be addressed. First, what moral basis should a self-driving car use to make such decisions? Secondly, how should this logic be conveyed in computer code? Researchers suggest using two ethical theories to program the behavior of self-driving cars: deontology and utilitarianism. The three laws of Azimov's robotics are a typical example of deontological ethics. According to this theory, a self-driving car must strictly follow the prescribed rules in any situation. According to utilitarianism, every decision made by a car should strive to maximize the usefulness of such a decision. In this case, it is necessary to define utility, one of which may be maximizing the number of saved human lives. According to researchers, self-driving cars should operate on the basis of a combination of several theories in order to be able to make morally sound decisions in emergency situations
2019 year
California allowed fully unmanned vehicles to deliver parcels
The head of Daimler said: to build a business in the field of robotic
Hyundai launches unmanned taxi in California November 4 with support from Pony.ai and Via
Scientists from the Georgia Institute of Technology, as a result of modeling, found that stopping only 20% of unmanned vehicles during rush hour completely paralyzes traffic in the city
In November 2017, a Navya unmanned bus knocked down a pedestrian in Austria in Vienna at a speed of 12 km / h. The victim received minor injuries
At the end of May 2019, Chinese startup Neolix Technologies began mass production of unmanned vehicles for the delivery of goods. Huawei buys them
2018 year
Unfortunately, there will be accidents. People will be injured, some of them will die, ”he said in an interview with BBC.
2017 year
Results are tabulated
2016 year
This conclusion allows us to conclude that Google has made significant progress in introducing unmanned vehicles [42], since the agency does not mind cars without drivers driving on highways subject to Google’s responsibility. But the decision of the agency cannot be considered federal law - this is only a recommendation.
2015 year
Google autocar accident. Google’s unmanned car got into a serious accident, which cost no casualties, but three Google employees were injured. A “self-propelled” Lexus RX 450h hybrid crossover drove up to the intersection when two cars moving in front of it suddenly began to brake, despite the traffic light allowing it to pass. The drivers of these cars saw a traffic jam after the intersection and, despite the burning "green", did not dare to go to the intersection of streets so as not to block traffic at the intersection. The standalone Lexus also braked. The driver of the car, driving behind the "google car", did not have time to react and crashed into a Lexus at a speed of 27 km / h. Later it turned out that the man driving this car did not even have time to press the brake pedal until the moment of impact. This accident for Google was the first in which people suffered.
почему тогда автопилотом БЕЗ пилотов нет на самолетах?
потому что они ценят свои бабки и не доверяют никаким алгоритмам управления
а для массово автопилота в авто и так сойдет )) считают корпорации
Comments
To leave a comment
The practical application of artificial intelligence
Terms: The practical application of artificial intelligence