From November 25th to 27th, the 2021 (5th) Gaogong Smart Car Annual Meeting was held in Shanghai. Zhang Dezhao, CEO of Smart Traveler Technology, was invited to attend and delivered a speech ON the theme of “Using Driverless Brain to Empower Various Industries to Build a Smart Life Circle”. The speech mainly focused on the definition of driverless driving, the characteristics and The commercial landing scene chooses three aspects to expand.
At the beginning of the speech, Zhang Dezhao introduced that Zhixingzhe is an autonomous driving company whose core business is the development and application of autonomous driving brains. At the same time, Smart Traveler has created a business model that applies driverless technology to multiple scenarios.
Autonomous driving shouldn’t just be unmanned
In the eyes of many people, driving a vehicle is a vehicle that can drive autonomously without a driver. But in Zhang Dezhao’s view, unmanned driving is more than that. He believes that once autonomous driving technology matures, the autonomous vehicle will no longer be a mere means of transportation and transportation, it will become a new life service platform. In the future, autonomous driving will be applied to various industries, such as common sanitation, logistics and distribution, as well as firefighting, border patrol, mining, port travel and other fields.
In the field of passenger cars, self-driving vehicles will become a new channel and a new channel in our lives. Specifically, in the future, self-driving vehicles will become a new platform that can integrate people’s lives, work, and entertainment. That is to say, we can not only take self-driving vehicles to designated places, but also work, meet and even entertain in the car. Eat and wait. At the same time, in the era of the Internet of Everything in the future, self-driving vehicles will also use their powerful brains to help people solve a variety of daily life problems. Hotel; For another example, after get off work, an autonomous vehicle will pick us up from get off work. If there are no eggs at home, the vehicle operator will obtain this information in advance and put an egg on the car, then we will most likely take the egg away of. Zhang Dezhao said that the driverless brain of Smart Traveler can build such a smart life circle.
Autonomous driving will bring us unlimited possibilities, and such a vision also lengthens the timeline for autonomous driving to mature. Therefore, Zhang Dezhao emphasized that autonomous driving is a marathon, and he has just run 5 kilometers.
Unusual driverless brain
Zhang Dezhao introduced that there has always been a process of mutual coupling and mutual optimization between algorithms and hardware, algorithms and computing power in the field of autonomous driving. Based on this, Smart Traveler has developed a software and hardware integrated brain that is different from most traditional unmanned driving brains in the industry. Autonomous driving solutions include AVOS software platform and AVCU hardware platform. Among them, the AVOS software platform is a full-stack autonomous driving algorithm from perception to decision-making to control; the AVCU platform is a hardware computing platform that supports AVOS, and is the industry’s unique hardware architecture that includes multi-domain controllers, which can ensure the efficiency of each module of autonomous driving implement.
Zhang Dezhao emphasized that after entering the era of artificial intelligence, data will be the key to the development of future autonomous driving technology, and will form the “triathlon” of the driverless brain together with algorithms and computing power. Based on this, Smart Travelers always insist on using a driverless brain to empower multiple scenes, including open structured scenes, semi-closed and semi-structured scenes, and three-dimensional off-road scenes. Massive data has been acquired through multi-scenario landing, which constitutes an automated driving data closed-loop system, creating a data-driven full-scenario, full-working automated driving mass production solution.
At the same time, in order to enable the driverless brain to be quickly deployed to various application scenarios and applications in various industries, Smart Traveler has also developed a full set of tool chains, including data mining, high-precision map tools, sensor markers, etc. In order to achieve data closed-loop, Smart Traveler has developed cloud Modules such as data annotation, model management and system simulation, thus forming a high cohesion and low coupling system architecture to ensure functional reuse and rapid deployment in different scenarios.
In terms of hardware, Zhixingzhe is also constantly improving, from the 3T version in 2018 to the 180T version in 2020. Zhang Dezhao revealed that the 500T version will be launched next year and will reach 1000T computing power by 2024.
In terms of data, Smart Travelers always believe that the core competitiveness of future autonomous driving will not lie in the algorithm itself, but in the cost and speed of the algorithm’s replacement, and this is based on the cost and speed of data acquisition. Therefore, Smart Travelers choose to use multiple scenarios to obtain data in richer scenarios at a lower cost and at a faster speed. At the same time, the actual commercialization also provides a better foundation for rapidly improving the ability to use data replacement algorithms, thereby forming a solid barrier to competition.
Multi-scenario landing is the best choice to maintain barriers to competition
At present, commercialization is the most concerned issue in the industry. Regarding how to choose the landing scene, Zhang Dezhao believes that a comprehensive consideration should be made from four aspects.
First of all, the technology is reachable. He said: “The standard for technical reachability is to remove the security personnel. As long as the security personnel are still needed, it will not be a real big commercialization.” The second is that the market is large enough. In Zhang Dezhao’s view, the complexity of each scenario, the overall investment, and the penetration rate achieved within a certain period of time are all comparable. Therefore, the larger the market, the more commercial returns will be brought. Just higher. The third is whether the social infrastructure is complete, such as laws and regulations, insurance, and public acceptance. The last question is whether driverless technology can really help this industry. Zhang Dezhao believes that every industry has a very deep know how. Our advantage lies in the autonomous driving technology itself, not the expertise of each industry. Therefore, Smart Traveler I hope to use driverless technology to empower traditional industries instead of subverting them.
In addition to the above four basic dimensions, Zhang Dezhao also pointed out that before realizing the vision of building a smart life circle with driverless technology, people’s life safety needs and the need to free people from repetitive and arduous tasks are two. Basic needs. Whether or not these two requirements can be met is also the key to the large-scale implementation of unmanned driving technology.
Based on this, Smart Travelers chose scenarios such as environmental sanitation, logistics, passenger cars and special vehicles for large-scale implementation. As of November 26, 2021, the application mileage of the Smart Traveler unmanned brain vehicle has reached 4.93 million kilometers.
If we follow the above criteria to choose the unmanned driving technology landing scene, we will find that there are many scenes that meet the above criteria. Whether to take a single-scenario route or a multi-scenario route is still a tangled issue for many companies.
As an advocate and practitioner of multi-scenario landing, Zhang Dezhao also gave his own views. He said, “Now that single-point demand has not reached the explosive moment, can multi-scenario landing not only meet the needs of data acquisition, but also maintain technical barriers. , And also build a complete, low-cost supply chain system.”
Finally, Zhang Dezhao emphasized that unmanned driving is a marathon, and it is only at the stage of 5 kilometers. There will be countless challenges in the future. Unmanned driving companies must have sufficient hematopoietic ability to ensure that they can run longer. Go farther.