(Original title: Fearing that unmanned cars are not safe? But 10% of the work will change as a result)
Qiu Zhili
“No matter how skeptical the existence of AI is, one or two years later, everyone has to accept the fact that the learning ability of machines far exceeds the ability of human learning.†Recently, the Udacity joint venture was initiated. People and President, Sebastian Thrun, former vice president of Google and co-founder of Google X Labs, emphasized in his speech to China.
He believes that in 2050, automatic driving will exceed human driving behavior. Future AI technology plays an important role in areas including cancer diagnosis and telemedicine.
But for traditional cars, this will undoubtedly bring about changes in the nature of cars, production methods and business models.
Whether it is the Internet builder or the traditional car manufacturer, first establishes a loop that can acquire data, enhance machine capabilities through machine learning of these data, and in turn improve the autopilot experience. In such a cycle, whoever has a large amount of data, who updates the data frequently, who is most likely to win.
2050 machine driving will exceed humans
Thrun joined Google in 2007 and was co-founder of the Google X and driverless car projects. He is also a professor of computer science and engineering at Stanford University and has already had important explorations in robotics and automotive technology. Now Udacity he is responsible for mainly teaches the world the technical knowledge of unmanned vehicles.
Why is AI faster than human learning?
In Thrun's view, AI learns faster than humans. Taking deep learning as an example, artificial intelligence, machine learning and human learning are completely different. “People may encounter traffic errors more or less. People hope that they will not make mistakes for the second time, but for humans, only personal experience will make mistakes. It will only take experience afterwards, but the same mistake will kill people to varying degrees."
Google’s artificial intelligence is input into the computer through the memory of driving functions, and then the computer learns, and through the continuous learning of the robot, all other people learn the functions and master the models of traffic accidents through the way of car networking. , so AI learns faster than humans.
Thrun firmly believes that in 2050, automatic driving will exceed human driving behavior. At the same time, with the birth and development of the concept of deep learning, AI technology has played an important role in areas including cancer diagnosis and telemedicine. With deep learning, technology will continue to advance. “If automatic driving really becomes a reality, about 10% of the work will change. For example, the job of a truck driver may be replaced. Whether it is a teenager or a centenarian, it is possible to achieve free driving. ."
According to Thrun, Google’s unmanned vehicle team and Uber have tested EZGO, an autopilot shared car project. Users can use an app to summon an autopilot shared car. Users can choose to drive the car manually. Let it drive automatically, and when it reaches the destination, the shared car can be automatically driven to the parking lot to solve the parking problem.
“Do we think of the car as a server? Or treat it as a private item? Is the customer wanting to own the car? Or does it only have the service the car brings him? This involves Business model," Thrun said.
Counterattacks of traditional cars
For the traditional automotive industry chain, this will undoubtedly bring about changes in the nature of automobiles, production methods and business models. "Autonomous driving technology will make the car a smart mobile space and fundamentally change the business model of the auto industry. People do not need to buy a car, but call an autopilot shared car," said Xu Heyi, chairman of Beijing Auto Group.
At present, the development of artificial intelligence is regarded as an important strategy for enhancing national competitiveness and safeguarding territorial security. The automobile field has always been closely watched as an AI key track. The situation in the global automotive industry is increasingly complex. The competition between traditional car companies is gradually shifting to the AI ​​field, and technology companies have always entered in a subversive manner.
“Since AI defines the car of the future, the future competition of the auto industry is AI talent. Such companies like BAIC are stronger than manufacturing, but they are facing AI challenges. New car companies such as Weilai Automobile have better technology and The genes of Internet companies are more likely to attract AI talent.†Li Bin, chairman of Weilai, responded by answering the question of which of the most popular car companies, BAIC and Weilai, was stronger. "So it seems that Weilai cars are more likely to win, just as Amazon defeated Wal-Mart."
Xu Heyi responded: "I agree with Li Bin in principle, but competition is not just about talent but also other resources."
From the perspective of technology landing, Audi no doubt walked in the forefront of the auto-pilot industry. However, according to Stefan Greiner, head of research and development of Audi's China Autopilot and Chassis, the auto driving heat has increased over the past few years. However, there are few car drivers out of car companies. Until two years ago, this phenomenon gradually changed. The mainframe factory realized that the autopilot technology was becoming more and more mature. It began to investigate autonomous driving and began to consider how to push related technologies to land quickly. From this perspective, car companies think more realistically.
“For the depot, safety is the first key element. It refers not only to infrastructure, but also to the entire sensor system of the vehicle. For example, self-driving cars have different radars, cameras, and laser systems. In the car, you need to verify its safety first,†Stefan Greiner said.
Thrun believes that the main engine plant plays an important role in the development of unmanned vehicles. After many years of building experience, the OEM is now faced with a whole new problem, that is how to better emphasize the importance of AI software in the manufacturing process to ensure that auto parts can adapt to autopilot technology after 20 years.
Still at the technical outburst
China's autonomous auto brands keep pace with leading international companies. However, the future development of autonomous driving still requires breakthroughs in laws, regulations, ethics, and infrastructure. At present, the safety of technology applications is a major pain point for consumers.
According to Chen Juhong, vice president of Tencent, the industry is now entering the first phase of technology outbreak. "In the next four years, 5G communications, new, constantly revised, and improved AI algorithms, low-cost laser radars, new client-side OS interfaces, and cloud-based technologies that will be supported by big data training will all develop explosively to stimulate commercialization. Artificial intelligence unmanned vehicles appear around 2021."
After the outbreak of technology, artificial intelligence unmanned vehicles are subject to battery energy density and other energy industry issues, basic transportation facilities needs, manual operation of vehicle life and other factors. Manually operated vehicles and commercial artificial intelligence unmanned vehicles will experience a 12 Around the year, the transition period was mixed, and finally it entered the third phase of the intelligent transportation era.
Today's auto AI evolution path still faces challenges of five dimensions: policy, regulations, infrastructure, high-precision maps, technical standards, and acceptance.
Wang Xia, president of the Automotive Industry Branch of the China Council for the Promotion of International Trade and chairman of the Chamber of Commerce of the China Chamber of International Chamber of Commerce of the Automotive Industry, expressed the same view. “The complexity of artificial intelligence determines that its development cannot be accomplished overnight, and now there are many automotive products that are beginning to It is equipped with a voice interactive system, but it does not have such a self-learning ability, and the functions of the car's physical conditions, our road conditions, people's needs, car networking, etc. have not been integrated into a single platform. Artificial intelligence is still in the early stages of development."
At the same time, the industry's large and small mergers and acquisitions and intensive investment boom have made smart cars an outlet, but many product technologies are similar and lack the core of original things. We must know that what we are facing is a technology-led change. If there is no core technology, blindly following the trend, even if it is blown up by the wind, it will hardly be a big deal.
USB Slip Rings are an increasingly popular way to provide power and data to rotating devices. By using a slip ring, you can avoid having to run cables through a hole in the device's enclosure, which can be difficult and messy. USB slip rings also provide a convenient way to update firmware or change settings on the device without having to open the enclosure. It is commonly used in applications where there is a need to send or receive data or power between a stationary object and a rotating object, such as in an industrial setting where a machine needs to be powered while it is still in use.
Oubaibo's USB Slip Ring, Rotary Swivel Joint manufacturer, chooses the high-quality Senring slip rings, and fiber Rotary Joint products. The company has been in business for many years and is a leading supplier of quality slip ring products. They offer a wide range of products, including high-quality sending slip rings, fiber rotary joints, and more. Their products are affordable and of the highest quality. They are a great supplier for those looking for quality slip ring products.
Usb Slip Ring,Rotary Swivel Joint,Senring Slip Rings,Fiber Rotary Joint
Dongguan Oubaibo Technology Co., Ltd. , https://www.sliprobs.com