Shulou(Shulou.com)06/02 Report--

Today's story is quite "strange."

In other words, many friends know that Tianjin is holding the "Third World Intelligence Conference" these two days. I was sorting out relevant topics and suddenly noticed such a message: Huawei Cloud City Agent announced its official launch in Tianjin at the "2019 International Smart City Innovation and Development Forum" in this conference.

What's special about this news for me is that I have written many times about Huawei Cloud's agents, and Tianjin is the first stop where "urban agents" landed. In addition, I live in Tianjin, which feels a bit like a chance meeting.

So I took a closer look, where is Tianjin's urban agent online? Then I found that the project had been launched in Tianjin Zhongxin Eco-city and deployed at eight major intersections of the eco-city.

The climax came-I found that the closest one to me in these 8 intersections was downstairs in my house...

This is probably the legendary, you do not see the scenery, the scenery will come to see you. I stopped downstairs to pick up a delivery and started an "interview" tour about the application of urban AI technology...

The closest field investigation to AI in history

Compared to all the interviews I have experienced, this "interview" does not cost any transportation costs and is low carbon and environmentally friendly. But I soon discovered that things were not that simple: although one intersection was close at hand, it was still a bit far to walk all eight intersections. After actually measuring, I checked all the intersections where Huawei Cloud City Agents were deployed, and walked more than 7 kilometers in total. If there are friends like me who want to find out, I hope this information can help you...

After identification, I soon discovered that the "physical appearance" of Huawei Cloud City Agent should grow like this:

Yes, it's this combination of traffic lights and cameras. Because after careful recollection, the common change of these eight intersections is to join this device. Let's take a look at the next intersection:

Then you can change the intersection:

In addition to the camera + traffic light device, another new device appearing at intersections is smart signs. There are many kinds of signs, and the following functions are to prompt pedestrians to cross the road:

Anyway, although these devices are not particularly "sci-fi," things are a little different when AI capabilities are added.

It should not be difficult to guess that the main ability of traffic light + camera +AI is to identify and understand vehicles on the road based on cameras, and then calculate through AI algorithm, and control traffic lights reversely according to AI operation results. In this way, we are familiar with the "car watching lights, counting seconds to pass" mode, into "lights watching cars, according to the number of cars released." At the same time, the information identified by the camera will also be displayed on roadside signs for traffic guidance.

This logic is not difficult to speculate today, but we can't stay at the level of knowing the "AI command" traffic lights. We also need to know more about Huawei Cloud City Agent, how to use it? What's the use?

Huawei Cloud City Agent, What is the mystery behind it?

In order to understand the mystery behind Huawei Cloud City Agent, I checked more information and consulted experts from Huawei Cloud and Eco-City respectively.

In general, the devices we see at each intersection are actually connected to Huawei cloud-based AI services as a whole. In the physical world, what we see are traffic lights, cameras, sensors and signs, and behind them are Huawei Cloud's edge computing system that provides real-time local computing power, and AI intelligence that performs complex AI operations in the cloud.

And when this system falls into real life, to be precise, it falls on my doorstep and becomes a three-dimensional ability.

I. Let AI command signal lights to ease traffic congestion.

We all know that the duration of traffic lights is fixed, and in many cases "waiting for traffic lights" is a waste of time. Probably no traffic at all at the green light, and a lot of traffic at the red light. This problem is often the first problem to be solved when AI enters urban scenes.

And letting AI command the signal lights is actually not that simple. For example, in Huawei Cloud City Agent landing in Eco-City, behind each camera, video perception capability and simulation prediction capability shall be implemented on the end side, and the collected vehicle information and traffic prediction information shall be uniformly delivered to "Cloud +AI" for processing. The cloud, on the other hand, deploys complex technology toolkits such as vehicle inference engines and real-time prediction engines, which will process the information collected by the cameras thoroughly and deliver it to the signal lights to start the corresponding red light early termination algorithm or green light extension algorithm.

After the integration of many AI recognition and reasoning capabilities, the signal lights can see the traffic flow and decide the timing.

In this way, the traffic lights at the main intersections in the eco-city will be able to adjust the traffic light duration according to the length of traffic.

For example, before joining Huawei Cloud City Agent, it is like this:

Here's how it works:

AI to determine the length of the signal time, this may be heard by everyone. The urban intelligence actually applied by Huawei Cloud this time also enables the signal lights to have many other capabilities.

For example, we often encounter such a situation: the green light has been green, but the horizontal vehicles fill the road and simply cannot pass. Wait a minute, the lights are red again. This phenomenon is called vehicle overflow in the field of traffic control. Because the green light to the red light does not change in time, the vehicle cannot pass, can only park the car in the middle of the intersection.

The solution to this problem is to be able to let the traffic lights on a road pass intelligently. Let one part pass and then stop the other exactly, so that there is no overflow. This requires the precise coordination of the signal system at the same intersection. Huawei Cloud City Agent of Eco-City already has this capability.

Before this ability was added, the beacon worked like this:

Now, with the help of AI, it has become like this:

In addition, Huawei Cloud City Agent also makes the signal lights of the eco-city learn various "divine operations" such as intelligent timing, main road and auxiliary road coordination, which are rarely seen in other intelligent transportation projects.

So, maybe the next question for everyone is, can such a complex AI operation really solve the problem?

The answer is really yes, because in the eco-city, whether the traffic jam situation improves or not, there is no need to look at complex data at all, this is a clear thing.

This is a traffic map of an eco-city, and as you can see, this place is basically an island. The main external traffic depends on the two bridges marked on the map. Because all traffic flows in and out of these two bridges, the main road under the bridge has become a daily "traffic feast."

A week after Huawei Cloud City Agents joined the Eco-City, the most congested places have indeed been significantly improved. I interviewed a traffic policeman in the eco-city who told me that he was on duty at the intersection of genial road and zhongxin avenue, where traffic jams and accidents were most common. With the addition of AI lights, the number of queue overflows in the morning rush hour has been reduced by 60%, which obviously reduces the probability of extreme congestion in all directions. And the morning peak of about 1 hour has been basically shortened by 10 to 15 minutes.

To confirm this, I decided to conduct an actual survey the next day during the morning rush hour. Friends who live in the eco-city know how tiring it is to block the Rainbow Bridge where you must pass every morning. However, after personal evaluation, compared with the past, the traffic jam situation has been reduced a lot, and the "long queue out of the island" at the morning peak has indeed been greatly relieved.

It is certainly gratifying that the morning rush hour can be greatly shortened. However, Huawei Cloud City Agents do more than that.

II. Use AI to open up "life channels" for special vehicles.

Based on AI multi-domain collaborative intelligent dispatching technology provided by Huawei Cloud, as well as real-time calculation of vehicle trajectory, GES graph calculation engine, traffic simulation and traffic prediction, urban agents have completed the ability to open green life channels for vehicles such as fire fighting and ambulance with one button. According to the test, using the green life channel, the speed of fire trucks arriving at a community in the eco-city dropped from more than 4 minutes to 2 and a half minutes, saving 44% of the time. There is no doubt that these times are equivalent to the safety of people's lives and property.

It is worth mentioning that the "green life channel" capability of urban agents is based on intelligent prediction of global traffic for early guidance. The traffic light duration will not be forcibly controlled, so that it will not affect everyone's normal travel.

Third, let AI bring intelligent travel experience.

Huawei Cloud City Agent not only brings changes in traffic lights, but also brings a lot of intelligent travel experience for vehicles and pedestrians through traffic display screens.

For example, through real-time monitoring and analysis of big road data, electronic signs can remind passing vehicles how long it may take to get somewhere, whether traffic jams are expected, whether there are construction sections, and suggestions to change roads, so that drivers have the ability of "clairvoyance."

In another scenario, urban agents can also help pedestrians crossing the road identify passing vehicles and even identify what cars are coming and alert them with voice. This makes the travel safety experience of the elderly, children and people in need of help very humanized.

Through the AI collaborative cloud, big data, edge computing and Internet of Things integrated solutions provided by Huawei Cloud, we finally see that the eco-city has gained three new traffic capabilities, which correspond to greatly reduced congestion, better security of residents 'lives and property, and a more humanized and intelligent travel experience.

There is no doubt that this collaboration has gained a lot from this place, but perhaps we can also ask a deeper question: After helping so much, what has AI gained?

Eco-city tour, what did AI harvest?

Personally, this interview gave me a good walk and a lot of knowledge. For AI technology, the harvest of eco-city travel may be richer.

In October last year, at Huawei HC Conference, Huawei Cloud released urban agents for urban scenarios. As mentioned above, this system is a fusion of multiple technologies, through the interaction of a large number of technologies and equipment, to achieve the digital world to the physical world of urban scenes.

Does that sound complicated? This is not your fault, when AI technology wants to enter such a huge scene as urban traffic, just look at a bunch of technical terms, it is difficult for ordinary people to accurately imagine what will happen in the end.

In traditional martial arts, sparring is not called a contest, but "mutual verification." I think Huawei Cloud's AI capabilities in the eco-city also happen to harvest this word: verification.

To this day, whether AI is useful in the real industry is still a controversial topic. The pros and cons of the topic each listed a bunch of truth. However, AI has never been a literary work, and there is no need to cause endless controversy. It's actually very simple, try it and you'll know.

However, Huawei Cloud chooses to land urban agents in eco-cities, which is equivalent to boarding the arena where the real chapter can be seen.

Tianjin friends may know better, eco-city has always been famous for introducing high-tech, especially intelligent technology, into the community. A large number of intelligent technology solutions are gathered here, and from one side, it also provides a suitable stage for urban agents. In less than ten days, we have reason to believe that Huawei Cloud City Agents have caught the move_morning rush hour is the best evaluation standard.

The value of City +AI has been verified.

In this cooperation, another AI harvest lies in that although only 8 intersections have been deployed, the special location planning of the eco-city basically covers the transportation hub of its whole territory. In some previous signal +AI projects, it was often small streets or areas that could cover experiments. The coverage of urban agents in eco-cities, to some extent, is the first time to open up AI collaborative traffic optimization attempts for the entire region.

In addition, Huawei Cloud mentioned when launching Tianjin Urban Agent that the six design concepts of urban agent are AI people-oriented, AI ubiquitous, AI always online, AI easy to expand, AI continuous evolution and AI step-by-step implementation. According to the consistent style of the eco-city, perhaps we will soon be able to see Huawei cloud city agents rapidly expand and evolve here, which will also be a very useful exploration for the city +AI proposition.

In short, AI today is looking for cities and industries that dare to taste the future. The good news is that the cloud has been found. Think about it from another angle, whether you can find good collaborators today, whether you can bring AI into the physical world, in fact, it also affects the success or failure of the cloud computing market.

Cloud and City: AI Main Battlefield in the Physical World

2019 can be regarded as a key year for AI to enter all walks of life and enter seven scenes of city, transportation, medical treatment, finance, education, manufacturing and retail. If the year before last was direction exploration and last year was conditional preparation, then this year's task is undoubtedly to launch a charge.

This charge will determine whether AI can survive, but also determine the voice of many technology markets, especially for the public cloud market.

We know that the traditional cloud service market is dominated by Internet, e-commerce and game enterprises, and their service demand is small and light, and the overall market stock has relatively clear limits. The public cloud, on the other hand, basically exists as an alternative solution for new enterprise IT services.

However, with the arrival of AI, the logic of the entire public cloud has changed rapidly. The value of AI is not business support, but to change the core production state and activate new productivity.

For example, the morning and evening peak of the eco-city ends 15 minutes early, which translates into half an hour's working time per person per day for these office workers.

Under this demand, large-scale government and enterprise departments, real economy industries and urban comprehensive formats will all enter the market demand sequence of public cloud. A whole new market is about to open up. The law of this market is that AI landing ability is equal to brand calling ability.

This is also the reason why Huawei Cloud is insisting on the challenge of "combining the digital world with the physical world" step by step.

Overall, from the cloud to the city, the AI main battlefield in the physical world has begun to become clearer and clearer. At least after I turned a few corners, I was impressed by the efficiency shift that AI brought in reality.

And the rules of the new battlefield are also challenging the stereotypes of the public cloud market:

1. AI landing based on public cloud, safety and reliability become the unshakable premise. And big factories with stability and security have begun to become the first choice for cloud computing.

AI began to enter the real world, testing not only AI capabilities, but also a combination of multiple capabilities. For example, in the actual implementation of Huawei Cloud City Agent, it involves many fields such as edge computing, Internet of Things, video cloud, etc., and technical comprehensive capability and engineering capability will become the new necessities of public cloud.

3. It is not difficult to find out from the case of eco-city that each city and each region has different characteristics for AI demand. However, industry AI is ever-changing. Whether it can understand individual needs and understand industry characteristics has become a new task for cloud manufacturers. In this case, ecological construction and developer construction have become urgent matters.

Layer by layer, we will find that behind a case, there is hidden the logic of AI changing urban life, the path of AI industry development, and the future trend of cloud computing market.

In fact, in the era of Cloud 2.0, all industries are actively using AI as an essential capability in the digital transformation phase. Facing all walks of life, Huawei Cloud is approaching various industries with its own technology, products and solutions, and landing "inclusive AI" with full-stack AI capability, so that all industries can truly realize AI"affordable, well used and assured."

Back to the question at the beginning, standing at an intersection, how far can you see the future?

Although I only went down one building, I really saw a lot, a lot.

(Part of the data in this article comes from the public number "Zhongxin Ecological City Release")

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.