11月27日下午，在2021亚洲青年领袖论坛“科技创新与智慧城市”主题论坛平行对话中，广州软件院常务副院长、中科智城董事长袁峰就“车路协同和新基建密不可分”的话题发表了自己的观点。

摘录如下：

回答第一个问题。其实，我觉得“新”呢，具体到我们这里就应该是新基建，包括主机厂和我们自动驾驶的团队，他们做得更多是单车上的事情。我们这两年经常在探讨自动驾驶，有两条主要的路线，一条是单车智能，一条是车路协同。车路协同跟新基建是密不可分，它一定是依赖于整个城市的基础设施。但是我们这些年也在研判，这两条路线一定在未来一段时间内长期并存，这其中也包括主机厂。比方说埃安的车，不能在南沙可以很方便地开，到了新疆不可以开了；小马智行的自动驾驶的服务，不能在南沙是OK的，到外地就不可以了。所以，我觉得单车智能还是会长期地发展并继续往前演进。

但是，在一定区域内，大家现在都在探讨这些问题，包括昨天我们在广州开的数字政府的峰会，其实就是智慧城市的论坛。我们讨论到智慧城市的基础设施、车路协同，包括红绿灯，包括我们讲的路侧单元。我们有很多路侧的基础设施，包括智能化的基础设施。我看前面那位专家提到这个智慧城市是integrated，所以一定是整合的。我们不管是哪个领域的工作，到最后要整合起来，为这个智慧城市里面的相关受众，也就是我们经常讲的车、路、人、云等各方面需提供一个整合的服务，它实际上是一个典型的基础设施的服务。我们讲车路协同的应用场景，在当前其实体现在很多的区域内。比方说园区物流车、扫地车、接驳车等等，包括像刚才莫总提到的端到端的高速公路的车，我觉得还是一个在有限环境下的具体使用，当然在未来我们相信一定会走向更广泛的应用场景。

我也简单介绍一下我们这个团队。我是科学院软件所的，于2011年从北京来到广州的。我们建设这个广州软件院，孵化了中科智城这样的企业，其实一直是配合政府在做新基建方面的一些工作，我觉得这是非常重要的。刚才上一个专家提到的整合的问题，这个问题一定是为我们的主机厂和小马智行这样的团队，以及我们在整个智能交通里面的各方面受众提供重要支撑的。它涉及到很多问题，包括政府怎么样在里面扮演关键的角色及我们各方面的制度，甚至包括我们的政策。因为小马智行的车现在在亦庄运行是要收费的，那么它一定会带来相关的监管问题，即法律监管、政策以及管理方面的问题。这些问题到最后，也是需要背后基础设施的支撑，所以，我们团队主要是在做这方面的一些工作。我们跟广汽之前也有很多的合作，包括在广州的我们的智能网联运营中心。我们跟张院长也在合作，小马智行也是这个生态里面重要的一部分。所以，我觉得在任何一个地方，包括在新加坡，在其他地方，政府以及这种基建第三方的角色也是非常重要的。我们希望可以多方合作，未来能够将智慧城市，包括智慧交通的生态运行得更好。

说到痛，其实我觉得有两个痛。第一个痛，就是我们经常在讲的非技术的，也就是管理模式和商业模式上的，这是智慧城市的一个常见问题。像刚才广汽的这位专家提到的5G的问题，我们经常讲基础设施，为什么会觉得5G的基站耗电多？就是因为应用场景不够。如果应用场景够了，它的耗电就不是问题，所以这是一个鸡生蛋蛋生鸡的问题。在智慧城市里面，我们广泛遇到这样的问题，也包括现在我们讲的车路协同的问题。车路协同一样会遇到这样的问题，我们的基础设施投入是非常重大的，而它的应用场景现在又不够多。但是，我们认为这是一个技术方向的未来，那就意味着是先投入基础设施的建设，然后才支持这种应用场景的应用。这里面就涉及到刚才说到的智慧城市是一个大的整合，它涉及到不同的受众（stakeholder），包括政府、企业和我们享受这些便利的市民以及各个方面。那么，政府在这里面扮演什么样的角色？也就是在这个商业模式里面基础设施的建设一般来说是需要政府的，政府是算大账的。

我们讲的非技术里面的第二块，就是商业模式的问题，也是我们在遇到智慧城市的挑战里经常讨论到的。那么，我们怎么样去设计一个好的商业模式，怎么样去设计政府的角色，政府要做到什么样的投入、什么样的责任、什么样的产出，企业和市民在这里面又得到什么样的收获，扮演什么样的角色，我觉得这是一个特别重要的命题。我稍微展开一些来说，比方说，之前特斯拉在郑州出了事情，驾驶员会说是由于我踩了刹车，但是你的车没有怎么样，而特斯拉也给出了自己的证据，但是这是主机厂自己给的，其实更应该由政府的第三方给出证据，我认为这就是一个监管模式的问题。在未来，像这些问题、这种挑战一定会越来越多，需要我们的政府、企业和我们各方面的用户在这里面扮演好自己的角色。

第二个痛，我就简单说，也是刚才莫总提到的。因为我们自己是做技术出身的团队，所以经常会遇到的问题是说“技术不是问题”，有时候我们想要申辩，技术其实也是问题。当我们的技术还没有达到一个理想的程度，比方说自动驾驶，我们的无人工干预的时间还不够长的时候，那我们就要循序渐进地、审慎地去推动。比方说我们在做自动驾驶的探索，一开始是有安全员的，安全员是必须在驾驶位的。而现在随着我们的技术提高，因为我们有百万公里的测试里程，所以现在这个驾驶位是可以没有驾驶员的。现在已经可以尝试收费。

那么，我们的政策管理也应是随着技术进步配套实现的。当我们的技术进步没达到这个阶段，其他的系统也没办法进行配套，所以技术也非常重要。但是，我们非常看好这个方向。最近住建部批准了国家六个城市的混行驾驶示范，其实也是依托我们的自动驾驶里程和技术的不断进步。所以，我们觉得技术也是问题。

总结来说，我觉得有上述这两方面的挑战和问题。


On 27 November, Yuan Feng, Executive Vice President of Guangzhou Institute of Software Application Technology and Chairman of Zhongke Software Intelligence, shared his views on the close relation between vehicle infrastructure integration (VII) and new infrastructure at the Technology Innovation and Smart City Forum of the Asia Youth Leaders Forum 2021.

The full speech is as below.

The word "new" means new infrastructure to me. Automakers and self-driving solution providers focus more on vehicles. In recent years, we've come to the realization that autonomous driving should be divided into two pathways: vehicle intelligence and vehicle infrastructure integration (VVI). Well, VVI is closely related to new infrastructure; it must depend on the infrastructure of the whole city.

However, during these past years, we've also realized that the two pathways will co-exist for a long time in the future. For vehicles such as Aion’s cars, it can’t be that they are able to run in Nansha District but unable in Xinjiang Province. The same goes for Pony.ai's self-driving services: it can’t be that it is available in Nansha District but unavailable in other regions. So, I think vehicle intelligence will continue developing and advancing for a long term. 

However, within a certain area, we have also been discussing smart city infrastructure and VVI. There are many roadside infrastructure – and intelligent roadside infrastructure – including traffic lights. An expert mentioned earlier that a smart city is an integrated city. I agree. Whatever field we are working in, everything should be integrated in the end. We should integrate relevant elements in a smart city, i.e., the often-talked-about cars, roads, people, and cloud. With every aspect integrated, we provide an integrated service. It's actually a typical infrastructure service. As for this infrastructure service, VVI is reflected in many areas, for example, logistics vehicles in industrial parks, sweepers, shuttle buses, and end-to-end expressways. In fact, these scenarios of VVI are still specific applications in a limited environment. Of course, I believe VVI will be applied more widely in the future. 

Allow me to briefly introduce you to our team. I am from the Institute of Software of Chinese Academy of Sciences (ISCAS). I came to Guangzhou from Beijing in 2011. We set up Guangzhou Institute of Software Application Technology and incubated enterprises like Zhongke Software Intelligence. We've been cooperating with the government on new infrastructure. I think this is very important. It is actually about integration mentioned earlier. It provides an important support for automakers, companies like Pony.ai, and other stakeholders in the entire smart transportation system. However, many issues need our attention, including how the government plays a key role in it – especially in regulations and policies. For example, Pony.ai provides paid rides in Beijing Economic-Technological Development Area. There will arise various regulatory issues, including legal issues, policy issues, and management issues. In the end, to address these issues will require infrastructure. This is what our team is working on. We've had a lot of cooperation with GAC before. In Guangzhou, we cooperate with Mr. Zhang from GAC Research Institute on our Intelligent Connected Vehicle Operation Center. Pony.ai is also an important part of this ecosystem. I think the government and third parties of infrastructure play an important role in any place, including Singapore. We hope we can be better run the smart city, including the ecosystem of smart transportation, through multi-party cooperation.

Regarding pain points, I think there are two. The first one is non-technical, which is management and business models. This is actually a common issue in smart city development. For instance, on the issue of 5G, why do we think 5G base stations consume a lot of power? It is because there are not enough application scenarios. If there are enough scenarios, we may think power consumption doesn't matter. It's a chicken-and-egg question. In smart city development, we often come across issues like this. This issue also applies to VVI. If we invest in infrastructure, which is costly, there are not enough application scenarios. However, because we think this is the future, we invest in the construction of infrastructure before considering the support for application scenarios. In terms of smart city, the smart city we talked about is a big integration. It involves different stakeholders, including the government, enterprises, and citizens who enjoy these conveniences. What kind of role does the government play here? In this business model, generally speaking, we need support from the government, which is capable of making long-term investment.

This is the second part of the non-technical pain: business model, which is an often-discussed challenge in building a smart city. How should we design a good business model: How should we design the role of the government? What kind of investment should the government make? What kind of responsibility should the government take? What kind of outcome should the government produce? What do enterprises and citizens gain from this? And what kind of role do they play? I think this is a particularly important proposition. I'm going to dive into it a little deeper. There was an accident involving a Tesla car in Zhengzhou. The driver claimed the car didn't stop after the brakes were hit. Then Tesla also presented its evidence. However, this evidence was produced by the automaker itself. This issue is about the regulatory mode. We think in the future there will be more and more of these challenges. Our government, enterprises, and users should each play our part.

I'll make it simple about the second pain point. As we are a tech team, we often hear people say that technology is not a problem. Sometimes, we would try to argue that technology is actually a problem. When technology has not reached an ideal level, for example, when autonomous driving without human intervention hasn’t been applied long enough, we have to push it step by step and with prudence. At the beginning, a human driver must be seated in the driver's seat. With technological development, however, since we have conducted tests covering millions of kilometers, we can remove the human driver. And we can try to charge fees now. Then, policy management should correspond with technological progress. When technological progress has not reached this stage, other management methods cannot be implemented in accordance. This is why we say technology is also very important. Actually, we are very optimistic about this. Recently, the Ministry of Housing and Urban-Rural Development of China approved mixed-driving on the road in six cities on a pilot basis. That's because self-driving technology is constantly improving, and self-driving mileage is ever accumulating. Therefore, we think technology is also an issue.