On October 28th, at the "Future Forum" awards ceremony and annual meeting, Li Kai, chairman of the board of directors and CEO of Innovation Forum, recounted the keynote speech. He said that the biggest obstacle to traffic in the future is humanity, and driverlessness will break this point.
In his speech, Kai-fu Lee said that if intelligent transportation can remove the human factor, it can increase the degree of security and make the entire traffic more predictable.
He believes that most of the causes of traffic accidents lie in people. People will have emotions, will be disoriented, or will not react quickly enough and will make mistakes.
Kai-Fu Lee mentioned that Google had done an experiment to allow more than 20 employees to drive their own cars back home. The employees were advised to sit in the driver's seat and be prepared to deal with possible dangerous situations.
However, Google found that employees rarely listen to this advice. They are 100% assured of driving to the car, and do whatever they want to do: watching maps, watching videos, lying in the back seat, playing with their girlfriends. affectionate……
This experiment allows Google to discover that people cannot be trusted. Kai-fu Lee also believes that driverlessness must be in place, and there is no such thing as man-machine collaboration.
Kai-fu Lee: Intelligent Transportation in the Age of Artificial Intelligence
I am very fortunate to be the director of the future forum and share with you the views on smart transportation.
Artificial intelligence is the most important technology in the world, and of course it will also be applied to smart transportation. We particularly expect that artificial intelligence will catalyze the progress of the transportation industry in the future. This is one of the most important areas for investment in innovation workshops. Of the companies we invest in, about seven or eight are doing this field.
The amount of data for smart traffic is one million times bigger than today
We believe that there are four waves of artificial intelligence coming.
The first wave is the intelligentization of the Internet. By calling a large amount of data on the Internet, Internet websites and APPs become smarter;
The second wave is business intelligence, which activates and generates commercial value for many years of data retained by many commercial companies;
The third wave is the intelligence of the physical world, which digitally uploads many of the past non-digital activities in the real world.
The fourth wave is a fully automated intelligent, unmanned, robotic, and other commercial applications.
Smart traffic involves mainly the third wave and the fourth wave.
Today's travel has various problems, from congestion to PM2.5, to safety accidents, fatigue driving, and so on. To solve these problems, we must have information flow to know in real time how many vehicles are running and where there may be congestion. .
Artificial intelligence is data-driven, and smart traffic is also. Internet companies such as Google and Baidu have a lot of data. Banks, insurance, etc. can also use existing data, but there are no data in many areas of the transportation industry, such as “where each vehicle is locatedâ€. The key to extracting these new data is to place more sensors.
The need for upgrading in the areas of retail, warehousing, logistics, education, healthcare, and agriculture will drive the popularity of sensors. As sensors become more and more popular, there are many basic things that will change, such as parallel computing and parallel computing. How to store one million times more data than today's data? A single click on Taobao today generates only a very small amount of data, but to know where a car goes every two seconds, the amount of data required is very large.
Mobai+Drops, making traffic predictions based on big data
Here I use several companies as examples to talk about the role of big data in smart transportation.
The Mobike we invest in is itself a smart sensor. Every time you use Mobike, finding, unlocking, riding, and collecting money will generate data. Using GPS, Bluetooth, and heat detectors, the driving information will be continuously transmitted to On the network, intelligent scheduling. How much data does this produce? It's amazing.
When Mobsee first saw China Mobile, it said that a car needs a SIM card, and when it said 7 million digits, China Mobile could hardly believe it. Now, China Mobile’s 1% SIM card is provided to Mobike.
A bicycle numbered No. from Mobike was put into use on April 21 last year. It was riding 1,180 times a year, traveled a distance of 2,623 km, and uploaded a message every two seconds. All Mobike bikes upload more than 20TB of information each day.
With more data, you can make accurate predictions and judgments. Traffic infrastructure is difficult to change, but smart scheduling can be used to improve traffic efficiency.
For example, the drop can tell you not only where the car is today; where there is a car after an hour; where it needs a car; but also where the future car will be and where someone will need the car.
Dr. Wang Jian did a demonstration at Alibaba. He picked a city and placed artificial intelligence in the city to make the traffic lights smarter. When a large number of people go to a place, give them permission, or an ambulance comes, let the ambulance go first.
The time and frequency of the traffic lights are modified according to the traffic flow through the vehicle. The results showed that in a crowded time period, the average traffic efficiency of the city increased by 15%, and the time for ambulances to arrive at the scene was shortened by 50%. This is a very simple form of future city brain and intelligent traffic.
The collected data is not only transmitted but also can be practically applied. For example, the application of computer vision in smart transportation, Face++ we invest in has a real deployment case, you can identify who is in the red light, capture the face, send the ticket directly to his home.
When two people stand on the sidewalk, with their hands on the side, it is more likely that they are chatting. The probability of entering the sidewalk suddenly is very low, but if two people lean forward, they prove that the probability of entering the sidewalk will be higher. The traveling vehicle makes accurate and safe judgments based on such predictions.
People can't be trusted, autopilot must be in one step
When does the driver come? The view of the innovation workshop is that the driver must be in place once and there is no so-called human-robot cooperative driving.
How to understand this sentence? Google has done an interesting experiment. They recruited a group of volunteers internally. Each person sent a self-driving car for test purposes and informed them that the car used for the test was not perfect. They still needed volunteers to sit in the driving position and were ready to respond to the car. Handling of road surface emergencies.
But Google found that volunteers rarely listen to this advice. Because in the vast majority of cases, Google’s self-driving cars perform very well and are free to deal with all kinds of complex situations on the road. In this way, almost every volunteer will be 100% assured to the driving operation to the car, he will use the time to travel, to do anything they want to do: There is a car when looking at the map, there is a car When watching a video, when you are riding in a car, you're lying in the back seat and you have a good time with your girlfriend.
This test gave Google a clear idea: Once the self-driving car has reached a high enough level, passengers in the car will take all the control right to the car. Regardless of whether the software of the self-driving car is still at risk at this time, regardless of whether the extreme road conditions on the road are properly handled by the self-driving car, the owner will not maintain 100% alertness.
In fact, people cannot be trusted. Google believes that to ensure the absolute safety of autonomous driving, we must not rely on the participation of the people. The driver must be in one step and be able to cope with all (at least very close to 100%) extreme road conditions. .
Automatic driving should be done on highways and trucks first
The increase in automatic driving requires a lot of data. How do you accumulate data in one step?
We believe that automatic driving must be limited in the scene. For example, you can first use the scenes, parking lots, airports, clean vehicles, garbage trucks, trucks, and other limited scenes to ensure safety, and then try auto-piloting of expressway trucks, and then do the driving scenarios of ordinary passenger cars and ordinary streets.
The trend-setting technology of innovative workplace investment is to make low-speed automatic driving of limited scenes without any danger. This simple scenario can also collect a large amount of data, and there is no security problem. The data is collected and iterated again.
On the road in Beijing, there is no person driving without a driver, because there will be a variety of unexpected phenomena, such as unpredictable children jumping out and traffic jams in the alley. But on the highway, the driver has opened better than people, because the highway is more controllable.
Trucks are actually very good driverless applications because there are no passengers inside and most of them fly on the highway and can collect data without danger.
You can also follow the method. The first truck is supervised. The rear car is completely unmanned. Just follow the car in front.
In the future, road rights may be subject to smart adjustments. For example, when they go to work, most of the cars go to the middle of the city. There are more cars that drive back to the suburbs when they leave work. The road rights will be adjusted according to the needs of people. Emergency vehicles, fire vehicles, etc. Get priority right.
Humanity is the biggest obstacle to smart transportation
In the future, AI will enter every car and make them truly unmanned. Cars can remind each other, puncture, to remind the car around to be careful; when the two cars go in the same direction, they come together to save energy. When they are separated, they will walk on different paths. This time will bring great changes to human travel.
However, drones still face many obstacles. What is the biggest obstacle? Look yourself to the left and to the right, human beings. In fact, most of the traffic accidents are caused by people. People will have emotions, they will feel bad, they will be distracted and they will make mistakes. When the driver is doing well, removing the person can not only increase the security but also make the whole traffic more predictable.
There are also people who are solidified. When the car first appeared, there was a law that required the car to be on the road. There must be one person holding a red flag in front of the car to lead the way, so that people who ride horses are scared. This is a very silly law.
In addition, drones face moral obstacles. In the United States, if a self-driving car hits someone, there is always a choice, whether it is a robot killing or how to solve the moral issue.
We need a very clear and effective law that will allow science and technology to advance without harming humanity. We must believe that a large number of data iterations will definitely give artificial intelligence a better driverless experience and bring smarter traffic. era.
I believe that in the next fifteen or twenty years, smart transportation will be the largest catalytic device in human society. When the power of capital, the strength of the industry, and top talents invest in this field, all challenges will be solved.
Features
â—† Small Compact Size,Global safety approvals
â—† Long life & high reliability.
â—† Wide Range of wires Terminals.
â—†Variety of actuator and terminals
â—†Customized Designs
â—†Widely used in household appliances, electric force, telecommunications, machine tool, vessel, textile, printing, Mining machine, Mustic Instrument, etc.
Miniature Micro Switch,Ultra Miniature Micro Switch,Micro Miniature Slide Switch,Miniature Snap Action Switch
Ningbo Jialin Electronics Co.,Ltd , https://www.donghai-switch.com