Text: Zheng Driving @ Tan Qing said AI editor-in-chief
On March 5, someone complained on Twitter, "Traffic lights should be controlled by AI, and it's crazy that we waste a lot of time with red lights at intersections where no one has no cars." ”
In less than a day, the friend was "flipped" by Musk, who replied to the tweet, "one hundred percent."
Not only that, but in the afternoon, Musk also expressed his own ideas for the solution to this complaint, "Maybe Tesla should make an AI visual device for traffic lights to maximize throughput based on actual traffic and traffic conditions." ”
Let the road end equipment and the car cooperate to build the future transportation, Musk's concept, can not help but think of many Chinese autonomous driving companies have been deeply cultivated for many years of vehicle-road collaboration solutions.
With Tesla, as the "catfish" of the electric vehicle industry, pointing directly at the new idea of vehicle-road collaboration, is vehicle-road collaboration really a "silver bullet" for Chinese companies to achieve overtaking in Tesla's autopilot curves?
Prospects: A contest between "building roads" and "going to the ground"
One of the significance of vehicle-to-road synergy is to alleviate efficiency problems such as traffic congestion in today's human beings, but to solve any pain points, the early solutions are often diversified. Yes, Musk, the leader of the revolution in the automobile industry, has long stood out on this issue.
Just like firmly taking the pure visual route on automatic driving and running counter to the world, Musk once again gave a "brain hole" answer to solving the problem of car congestion: digging underground tunnels and making underground tunnels into a vacuum.
To this end, Musk also established a company, but from the name of the company, Musk should be quite Buddhist for this plan - The Broing Company (boring company).
For his own tunneling plan, Musk once said in a TED speech, "The cost of building a tunnel in Los Angeles is about $1 billion a mile, so we must reduce the construction cost by more than ten times." ”
For Musk's concept, some people have used the reverse idea to challenge it, that is, to build a "flying car", but compared with the imaginative "heaven and earth", the vehicle-road collaboration that many Chinese technology companies are climbing the peak seems to give us a more practical and feasible new idea: through the technical solution of vehicle-road collaboration, to achieve intelligent transformation of existing transportation infrastructure.
In fact, due to the unique advantages of China's transportation infrastructure efficiency and other aspects, in recent years, this scheme camp has gradually condensed many Chinese enterprises, and vehicle-road coordination seems to be becoming a unique "Chinese solution" that challenges the end of automatic driving.
Different from Musk's tunnel concept that simply wants to improve traffic efficiency, in the view of Baidu founder Robin Li, the intelligent transportation created through vehicle-road collaboration has three problems to be solved, namely "safety, efficiency and low carbon".
"We always believe that vehicle-road coordination is a necessary condition for making autonomous driving land faster." Xidi Intelligent Driving, which focuses on the automatic driving of commercial vehicles and vehicle-to-road collaboration technology, said to Tan Qing that AI.
At the beginning of its establishment, Xidi Intelligent Driving had made strategic deductions, and judged in 2017 that the landing of commercial vehicle autonomous driving would precede passenger cars.
On the other hand, bicycle intelligence may soon reach more than 99% of the automatic driving capability, but for the corner case after 99.9999%, it needs to be solved by the cooperation of vehicle-road coordination.
Therefore, From the very beginning, Siddi Intelligent Driving has determined the "two-legged walking" model of "commercial vehicle automatic driving + vehicle-road coordination".
From the perspective of industry development, the current large-scale commercial landing of autonomous commercial vehicles in mines, ports, airports, and parks, as well as the vehicle-road collaboration projects arising from new infrastructure, have blossomed across the country, which seems to verify the correctness of the choice of Xidi intelligent driving routes.
However, Chang Bin (pseudonym), an institutional investor who has participated in investment in several autonomous driving-related companies, told us:
"Take the largest passenger car in the market, this is actually a very tolerant market for technological transitions, such as assisted driving technology, but it is not so simple to accommodate vehicle-road collaboration, and the early commercialization of vehicle-road collaboration may be more 'small and beautiful', and in the long run, it may be a bit like doing Robotaxi, the prospects are very good, but the future challenges are still very large."
As Chang Bin said, although the prospects of vehicle-road synergy are beautiful, what objective challenges do they face if they want to reach the end? We might as well analyze it further.
Challenge: "Car Lights" and "Street Lights" in the Finale
Pan Jun, CEO of Nebulas Interconnection, a Start-up Company of Vehicle-Road Collaboration that recently received nearly RMB200 million in Series B financing, believes that "building a globally interconnected intelligent transportation system of 'people-vehicle-road-cloud-network' is a new ecology for future transportation construction." ”
For the current stage of vehicle-road collaboration, this actually points to a technical problem, "people, vehicles, roads, clouds, networks", for mature vehicle-road collaboration programs, the basic elements can be described as indispensable, the most basic "car", the importance of which is self-evident.
Therefore, Tan Qing said that AI believes that the two schemes of vehicle-road collaboration and bicycle intelligence seem to be opposites, but in fact, they are the relationship between inclusion and inclusion to a certain extent. Creating a smart enough car seems to be the essence of the bicycle intelligent route, but it is actually one of the basic needs to ensure the smooth implementation of the vehicle-road coordination program.
At present, some of the head autonomous driving companies that take the bicycle intelligent solution are conducting road tests in California, USA, and from the current car collision report records on the official website of the California DMV (Motor Vehicle Administration), as of March 4, 2022, DMV has received up to 426 autonomous vehicle collision reports.
Not only that, but in these reports, there are still a large number of accidents that human drivers can completely avoid, such as the accident that occurred in Xiaoma Zhixing last October, and its test vehicle directly hit the traffic sign when changing lanes.
This seems to mean that the current level of intelligence of self-driving cars still needs to be marked with a question mark.
Although many autonomous driving technology companies such as Baidu, Xidi Intelligent Driving, and Mushroom Car Union are betting on the coordinated route of vehicles and roads, for the start-up industry, an intuitive phenomenon of healthy development is that the industry can accommodate diversified technical paths. In the face of vehicle-road collaboration, there are actually many different voices in the industry.
Han Xu, founder and CEO of Wenyuan Zhixing, once mentioned a point of view on vehicle-road collaboration, "The relationship between vehicle-road collaboration and bicycle intelligence is like the relationship between street lights and lamps. ”
In Han Xu's view, with street lights, the car will drive better and safer; but the car must have lights, such as suburbs, such as road end facilities problems, the car needs to provide its own lighting.
Therefore, the safety of vehicle-to-road collaboration vehicles will be improved, but bicycle intelligence must be there, after all, some problems cannot be solved: first, the cost of full configuration of vehicle-road collaboration facilities is very high; second, there are many areas that cannot fully install these facilities for various reasons.
Wenyuan Zhixing also made a simple mathematical model, 10 kilometers of road, 200 meters of a street lamp pole, a total of 50 street light poles, if you rely entirely on the road end sensor to achieve unmanned driving, a street light pole in 365 days If one day is not bright, 364/365 and then do 50 to the power, this probability is 87%, that is to say, unmanned driving ten times there is a time to hang up.
It is not difficult to find that in the problem of realizing the end of automatic driving, although it is to solve the problem of 0.001% left by bicycle intelligence, there are still many problems such as cost and late maintenance complexity under ideal circumstances.
From the perspective of capital, Chang Bin expressed to us another aspect of his hidden worries, that is, the current car-road synergy track is still surrounded by technology giants such as Baidu and Huawei, and under the very large industrial synergy needs of smart transportation, the resources and capital advantages of the giants will be more obvious, so that there is a risk of squeezing the market share of start-ups to a certain extent.
We can't deny that there are still many challenges in the face of solving the end of automatic driving, but we also can't ignore Musk's "awakening" to the problem of traffic jams, and the car-road coordination camp has continued to grow in China in recent years.
Combined with Chang Bin's hidden worries, Tan Qing said that AI believes that a very big problem in the current vehicle-road collaboration scheme is actually whether startups can have the vitality to maintain long-distance running under the "deterrence" of giants, which also poses a challenge to their commercial landing ability in the short term.
Outlook: The opportunity of "laying eggs along the way" of vehicle-road coordination
Although the Vehicle-Road Synergy Scheme is still in a state of chaos to a certain extent, there is still a long way to go from the end, but this does not seem to have stopped the recent investment enthusiasm of capital for the Start-ups of Vehicle-Road Collaboration.
According to the Tianyancha APP, last year alone, Xidi Intelligent Driving obtained 4 rounds of financing; Mushroom Car Union received strategic financing from Tencent and JD.com in September last year; and Nebulas Internet has obtained 6 rounds of financing since 18 years.
We have already mentioned baidu, Xidi intelligent driving, mushroom car federation and other enterprises for the recognition of the value of vehicle-road collaboration, so what kind of way can the value of vehicle-road collaboration be released quickly?
An insider of an autonomous driving unicorn company once told us, "In the future, we will definitely be related to diversification, this correlation is not only technology-related, but also industry-related, industry-related that is, we will walk along the semi-open road such as cargo transportation, and will not consider the problem of transporting people." ”
Although there is still a lot of room for development to be made up for by vehicle-road collaboration to achieve the final regret of filling the 0.001% style of bicycle intelligence, the current active bus priority system with the help of multi-scenario opportunities, such as Xidi intelligent driving and other enterprises, has given it a certain degree of commercial landing ability today.
The automatic driving industry often talks about "climbing Mount Everest and laying eggs along the way", Tan Qing said that AI believes that although the vehicle-road collaboration has now been discovered by many companies, it is undeniable that we are still not short distance from the day of the real national automatic driving, which seems to mean that the enterprises in the track need to "lay eggs along the way" to maintain the commercial vitality before climbing.
Li Hengyu, vice president of Xiaoma Zhixing, has publicly stated that Xiaoma Zhixing can use 80% of the passenger car autonomous driving technology to trucks. This means that in the bicycle intelligent solution, when grabbing the new market across scenarios, enterprises can reduce the input costs of time and capital and technology through technical reuse.
This means that when the vehicle-road collaborative track climbs to the peak towards the end of autonomous driving, it chooses a relatively fixed scene such as mining areas, buses, trunk logistics and other lines in the early stage, which is an excellent way for start-ups to continue to revitalize new technologies and land successfully.
In general, at the moment when capital is gradually returning to the rationality of the autonomous driving industry, for the start-ups of chelu collaboration, before rushing to the sea of stars envisioned by Robin Li, a game between survival and development has been quietly staged.
With the continuous evolution in the pursuit of survival, we also sincerely hope that the "Chinese plan" of vehicle-road coordination can lead the mainland to achieve the curve overtaking of the autonomous driving technology of Western scientific and technological powers in the near future.