Give some unmanned views 2020-08-0 1

First of all, I declare that I am an amateur, so let me express my opinion, not on technical details, because I don't know technology either. Don't spray if you don't like it, and don't despise it.

Many of you are knowledgeable professionals and may have different understandings of unmanned driving. I'll just talk about my trivial personal considerations and consider it a science fiction theme.

Let me talk about the current road environment first. The road driving scene is too complicated, external roads and internal roads, expressways and ordinary roads. As long as there is actual human participation on the road, there can be no great development of unmanned driving, because human driving habits are really strange, which is related to human cognitive level. When a driverless car meets a manned car on the road, it will be badly beaten. Someone is overtaking, not getting off the bus, or suddenly stopping, changing lanes without lighting or wrong lights, stopping at the side of the road without lighting, and the turtle changes lanes frequently. I think your driverless car can react? Don't tell me to limit the unmanned driving speed to a very low level and leave enough distance to react. In addition, as long as there are people on the road, our technical feasibility is bound to be restricted by ethics and law.

Unified standard road surface estimation with excellent infrastructure can be realized, but pedestrians and non-motor vehicles must also be completely separated from the driving road, without complicated intersections and too many kinds of traffic signs. It is best to separate unmanned and manned roads, that is, the road conditions are simple and single. This can greatly reduce the complexity of the algorithm and avoid the constraints of ethics and law.

I don't know if you have noticed that sometimes at an intersection (usually the intersection of the road under the viaduct), the traffic light at the intersection where the vehicle is close is actually not too far away from the next traffic light (maybe the distance between the two traffic lights is less than 100 meter, or even nearly 30 meters). For the current identification system, can it distinguish whether this traffic light is at a far intersection or near an intersection?

In addition, can the current unmanned identification technology distinguish the traffic lights at intersections from the taillights of other vehicles (some taillights are round)?

In fact, there should be too many aspects to consider, which will undoubtedly increase the difficulty of the algorithm by several orders of magnitude.

For the current human cognition, this form of unmanned driving is technically difficult to realize. Secondly, even if it is realized, is it really of great value?

My simple cognition is that one day, unmanned robot cars will replace human travel, and human beings will gain experience through VR simulation travel. Humans hardly need to drive their own cars or appear in them, but they can solve it by sending robot cars in their daily work.

There must be questions.

Q: If the roads are completely closed, it is better to vigorously develop rail transit. Rail transit is easier to be fully enclosed than roads, and the signal system has long been the so-called car networking. One day robots can solve all problems without going out, which is as promising as interstellar migration, but out of reach.

A:

"What do you want to achieve with unmanned driving now? I feel that it is nothing more than my own car (not public transportation) to take me out, put my life in this driverless car, be carefree, and fully trust driverless. But what's the difference between this practice and hiring a full-time driver? What's more, there are various driving behaviors on the road at present, such as parking at random, turning off the lights before parking, driving at turtle speed, changing lanes at will, changing lanes without lights, forcibly overtaking, drunk driving, speeding, fatigue driving, electric cars and pedestrians. What advantages do you think unmanned driving has over traditional human driving on the road? If there are too many personalized needs of human beings, autonomous driving will be too boring, and traveling with you by bus will also be boring. Driverless technology is worth studying and developing, but no matter how awesome it is, it looks like a Sabi AI under the current road conditions.

What's more, due to the lack of technology at present, it is difficult to completely synchronize the precise positioning system (GPS or Beidou) with the speedometer, because there is a difference between GPS speed measurement and traffic police speed measurement. For this kind of defect, the driverless car may not be able to quickly make extreme operations at critical moments to avoid danger, but can only stop the car or react far away from the known danger, which means that many technologies are not sensitive and accurate enough at present.

As long as there are people on the road, this kind of unmanned driving is bound to be restricted by traditional moral laws and bioethics. But this technology still needs research and development, because it is likely to be useful in other fields. However, with the technological breakthroughs in more fields, and more importantly, the great technological leap brought about by the fundamental change of human ethical and moral foundation will be realized. To a great extent, human ethical cognition defines the ceiling of many technological developments.

So will the moral law of interpersonal relationship change fundamentally?

I think so, yes!

With the popularity of mobile Internet, the foundation of human relations has been gradually weakened. The new mobile Internet will promote the formation of Internet-based relationships, and the foundation of this relationship will gradually break away from the influence of consanguinity. The weakness of consanguinity will inevitably gradually disintegrate human traditional ethics and current laws. In addition, after the emergence and popularization of artificial uterus, the relationship formed by human consanguinity will gradually decrease until it disappears, and the current mode of reproducing offspring by family will be replaced by the way of mass production and raising of offspring by artificial uterus factories. "

Others believe that,

Q: Didn't people hold back at home during the epidemic? Most people don't want to stay at home all day except the real otaku and otaku.

Autopilot can't adapt to the real scene, but let the real scene accommodate autopilot, which is a bit of putting the cart before the horse.

If all roads are really closed, it will have a negative impact on the city appearance and pedestrians/cyclists. ""

For the first point, I really dare not say that the driving distance is basically more than 3 kilometers. If you run around your home, take a walk and go to the supermarket to buy something, these activities within three kilometers are also outdoor activities. I didn't say I should stay at home completely, which is different from squatting at home during the epidemic.

Second, isn't it just to make the environment more suitable for human activities that human beings constantly transform the face of cities and even the face of the earth? This allows the real scene to cooperate with autonomous driving. What is the significance of developing unmanned autonomous driving for field use? Do you put humans in cars as rats? Why does someone have to be in the car? Do people in the car drive automatically just to avoid driving by themselves? You will still be tired by car. If there is something wrong with the car or there is an accident, can it guarantee that human beings will not be hurt? In this way, this form of autonomous driving is unnecessary for human beings. If you have to talk about disabled people, people who can't drive can go out by themselves without driving. Then I want to say that the role and significance of driverless driving are too narrow.

Thirdly, regarding the negative impact of closed roads on human non-motorized travel, most of the problems are only from the current time dimension. Simply looking at the single unmanned technology may consider factors such as pedestrians, but the unmanned technology is really mature, and many other high-tech technologies will certainly appear. These high-tech technologies have changed people's way of life, but also completely changed people's travel habits and ways. Due to the technology mastered by unmanned driving at present, the current road regulations, traditional ethics and human cognitive level, unmanned driving can not be widely used even if it is technically realized. In order to truly popularize the use of driverless cars, there must be no pedestrians on the road, and all driverless cars can have passengers or not. Don't say that all these conditions are met, even if any of them is too difficult to achieve. In the final analysis, it is impossible to achieve unmanned driving at present, but the technology must be developed and invested, because it is likely to be used in other fields.

However, I prefer the scheme of unmanned robot cars with no one involved in road traffic at all: unmanned robot cars replace human travel, and human beings gain experience through VR simulation travel. Humans hardly need to drive a car or appear in a car by themselves. Daily work can be solved by sending robot cars. ""