 Hi, everyone. This is my honor to share some considerations about the communication and connectivity for autonomous vehicles. In my point of view, I think the connectivity brings some new characters to autonomous driving. First of all, the collection based on cellular network makes the system possible to keep growing by the software update and the data collection. And the second one, the second character, the cooperation between vehicle and infrastructures extends the service to the vehicle and even make the autonomous vehicle more intelligent. The collection may via cellular communication or CV to X communication. And the third one is enhanced safety by cooperation between vehicles, definitely via the robust and the low latency dedicated communication. Then I would like to share some view or points about these new characters and some practice of China automobile in the following slides. The first character introduced by the community is growth. There are some reasons and the motivations. First of all, it's one of the user's requirements. The users, especially the young generation customers, they expect to have the same experience as smartphone in the intelligent autonomous vehicle. They expect the software can update it. They expect the post purchase feature enhancement is available. They expect the system become more and more robust. This is the one reason. And the second reason, the second motivation is from the OEM. The OEMs want to create more values from the same hardware platform, which is expensive with big computing power. Take talents car as example. The price of new models range from 18,000 US dollars to 23,000 US dollars. And the level 2 plus autonomous vehicle will take around 10% to 20% of the cost. So we are trying to create the technical possibility to add more features based on the same cameras, based on the same domain controllers. We are exploring some new business models to earn money from this system. So when one of the systems can grow, we can earn money from these new features. The third reason, in my opinion, is a high-level autonomous technology is not that mature. Now, we even say that it's a high-level autonomous driving system, but it's still a kind of driving assistance, very advanced driving assistance system in the vehicle. So this technology will take the time and the date to be mature. So this is the reason we think why we want this vehicle system keeping growing up. And to support this type of autonomous driving system, of course, the software can be updated. The HD map or ADAS map can be updated. And the configuration information can be loaded into ECUs. What special thing for autonomous driving system is that there are a few associated ECUs in this system. Between these ECUs, the signals are exchanged. The new feature is enhanced or updated. The associated ECUs should be updated in the same time. To be more specific, when you enhance the autonomous driving features over the air, you update the software of the autonomous driving domain controller, the smart sensors, the gateway, the infotainment system, the instrument cluster in the same time. So this makes the system level solution very complicated and also the size of the software package is large. Furthermore, in terms of the development methods for an autonomous driving system, data-driven development is the trend. The data of the EDR, the vehicle working driving habits, raw data of the sensors, and etc. is collected to develop new features to improve the performance of the existing autonomous system to validate the system. According to our estimation for our next generation autonomous vehicle, in the early stage that around 10 gigabytes of data will be collected to the dot. So this is the amazing size of the volume of the data. So we are thinking to achieve such kind of feature that low latency, high bandwidth, signal communication is a foundation and amazing data volume should be supported by new business models. Next slide, please. So according to China's practice, there are some points I would like to share with you. First, our first step is to want to are all of our vehicles to be collected. So by 2020, all China's vehicles are leveraged, so longer produced, longer collected vehicles. This is the first point. And the second one, we found that the external network like the 4G, 5G can meet our requirements. However, our bottleneck is the internal network and the E architecture. Why? Because we, in our build models, we keep a few variants of each build models, the low-level, mid-level, high-level, and even more variants. So the E architecture is kind of distributed and some kind of integrated type, but not the centralised or zonal type. So it's more challenging to have the vehicle-wise software updated or data collection. Next slide, please. Then the second character is a service extension based on the V2I. So which scenarios should be our point is that the more complicated the road conditions are, the more V2I-based service is limited. Why? It's because that we were thinking that in the simple scenarios, the direct, the active sensors can make a good effect. But in the complicated scenarios, the communication can help a lot. So in our point of view, we think that we have better deployed the V2I service to parking areas first, then to urban, then to highway. This is also the point that we propose to the Chinese government. Now, Chinese government have been promoting a few kind of projects on highway and the city area already. For example, China's longest and the most complicated intelligent highway will be complicated in 2021. It is around 128 kilometres and 47% are terminals and bridges. And there will be more than 300 ISU will be equipped. And there are around 10, more than 10 cities in China. They have the demonstration areas for the smart road and for the text for the collectivity car. And a few interpoints had some demos in parking areas to provide the demos like the reservation for parking lot or the map distribution. So this is the kind of data in China. And what China have done, first is that we build up our smart campers and build up, we just modify our campers to smart campers and the parking slots in our R&D center to be smart, to be collected. They have more than 80 smart ISU has been equipped in our campers and the parking slots. And we join in the demonstration activities of a four-level road intelligent vehicles and other demonstrations in China. Let's select please. The third character is enhanced safety by cooperation. In our concept events, if we want to talk about if the autonomous vehicle, the high-level autonomous vehicle, can drive better than human being, that the communication is indispensable. Because it has the capability to sense the long line of sight scenarios. So if we were aiming at the autonomous driving in all conditions, the communication is necessary. According to a publication from WIMO, now we can see a few collision and minor contact events in automated driving tests are published. So in this slide, there are three examples here. We can see that all these scenarios can be really covered by the sensors like LIDAS cameras are used on WIMO's automated car. Even it's already a kind of high level autonomous vehicle. So we still think that if we want to get this kind of enhanced safety, with the communication, the V2V application should be deployed. The next slide please. Because the V2V standard has not been really released in China, so in the past 10 years, China led or participate a few of the Chinese government funding projects trying to have to get some research, have some research and advantage activities in correctly automated vehicle areas. And we also joined and lead some of the to establishing some Chinese national industry standards. And we also have developed the 13 V2V applications in-house. This is what we have done in the past. Next slide. This slide is also the last slide. So when we're talking about these three characters, we want to know when we can see those new characters in production vehicles. So my point of view is that I'm thinking that we can see the character of growth very soon because the OEMs have enough motivations. They will make their car to be very open to be collected and have a better e-architecture. It will be a very popular character in most of the new production cars in in the next two years. This is my point of view. And the full vehicle eye service, I think this is leading or driven by the government at this moment, like in China. So because the Chinese government, they want to advance its transportation to digitalization and the use of the artificial intelligence, and the government would like to invest for the infrastructure. So my estimation is that around 2023 to 2025 that we will see some real production in the market. But for the V2V, the enhanced safety, my point of view is not that optimistic because it will be driven by the industry. So I'm thinking that if we will see the real production, the massive production that will be after 2025, that's my view. Okay, that's all the points I would like to share with you. Thank you.