Whether the story is among the ages of trickery, military war, or large business, sending its own personnel to the “enemy” undercover to steal secrets always frightened but most spectators feel exciting plot. In recent years, this classic plot staged frequently in the field of autonomous driving, which mixed with the exodus, betrayal, faith substandard many, what is the human behind the “twist”, or moral “bankruptcy”? This time we find out.
you that technology is good, the next second I was up in January 2017, Tesla will Autopilot former project director of Sterling Anderson to court, saying that at least 12 attempts to poach engineers autopilot, while also from the company took several hundred GB capacity of confidential proprietary data; February 2017, the parent company of Google and Uber Alphabet automatic car driving department to court, accusing Uber stole trade secrets to start their own Uber automatic pilot project ; March 2017, the security Division of Baidu, general manager Ma Jie said, there are a large number of “hackers for hire” is attempting to steal Baidu autopilot technology; July 2018, Apple filed court documents to the Northern District of California against Apple former employee theft of trade secrets; March 2019, Tesla its former employees, current employees CAO Guang Zhi car Xiaopeng file a civil lawsuit for allegedly stealing Tesla Autopilot technology source code. …… so frequent incidents of theft, if imitation confirms that piece, “This technology is good you can be the next second will fall to me.” Autopilot less “technology theft”, but also Dezhuan it? After the event we will explain in detail the analysis found that even technology theft is often divided into a variety of situations: one is the executive involved in the project / core technical staff due to the company’s management philosophy with the original or not, choose to leave home business establishment of the new company, resigned after the company took before confidential documents. For example, Google sued the parent company Alphabet Uber stole own trade secrets is a typical example. 2016 Uber acquired a company called Otto, the company founder Anthony Levandowski (Anthony Levante Duo Siji) worked in the automotive sector Alphabet autopilot software engineer. The biggest dispute between him and the Alphabet is on automaticDriving test, Google believes autopilot experiments need to experiment on the basis of security, which is the human need driver sitting in autonomous vehicles driving position to take over at any time. The Levante Duo Siji think, “If your job is to improve the technology, then security is definitely not your primary consideration.” Later, Levante Duo Siji and took a different view of Taylor sat a self-driving car departure, en route the car can not handle the situation encountered with a friction UIC Camry, parked in the center of the highway, and also lead to Taylor spinal injuries. Google does not satisfy the ultimate subject of Levante Duo Siji want to create a new company autopilot was expelled from Google, before leaving downloaded 1.4 million documents have become technology theft of irrefutable evidence. Another technique theft case is employee turnover, job-hopping behavior before, for example, the aforementioned Apple, Tesla former employee accused of stealing technology are attached to this category. Through public information shows that in this case, former employees often acted unprofessionally on the eve of departure, such as his network access behavior and activity compared to the time before the work of exponential growth, and will visit some of the original do not belong the content of his scope of work, and even download files to transfer data to the core secrets relatives of electronic equipment. In summary, the reasons for these people steal technology often out, and temptation and other original company philosophy substandard want to realize their ideas, quit. However, the parties to put aside their own positions and ideas from throughout the autopilot industry point of view, frequent incidents of theft technology and the industry’s own particularity and behind the control of resources, closely related to technical barriers.
automated driving technology behind the theft of fact, when we talk about automatic driving skill frequent incidents of theft when, in essence, is to talk about the team personnel, technical barriers and bottlenecks in industry and other fields autopilot problems. 1, human resources personnel China Automotive Research Association, executive vice chairman and secretary Zhuming Rong also mentioned that, according to incomplete statistics, China’s auto industry personnel functions less than 2 million of the total, while the automotive industry due to the lack of talent with network functions, serious It restricted the development of China’s auto intelligent networking. This phenomenon not only in China, the global auto industry driving the same. International Journal of conference papers, AMiner on autopilot TOP1000 scholars ranked the world’s h-index calculation and analysis are drawn out of the distribution map shows that the number of US personnel in the field is China’s 10Times or so. Nevertheless, the US company autopilot same lack of talent, big companies have joined the “grab Wars.” According to foreign media reports, Pittsburgh autonomous vehicles market is extremely fierce competition for talent, Uber, Ford’s Argo AI and other companies are all making great efforts to “grab” the relevant professional college graduates will be able to get $ 200,000 (about US 1.309 million yuan) in compensation. Based on this situation CONTROL ENGINEERING China Copyright , R & D centers in the United States as a good way to deal with domestic enterprises. According to incomplete statistics, as of now, Foxconn, Baidu, Wei to car, Ranger vehicles, plus wisdom science and technology, Tucson future, pony Chi and other related companies have R & D centers in the United States. Specific to the functional teams responsible is to provide the core technology from the United States team, the team responsible for landing the domestic R & D results US team in the country. Wu Zhou, vice president of automotive autopilot Xiaopeng introduced, autopilot Sino-US team of the division of general core algorithm team in the United States, they are responsible for providing the core technology, to complete the optimization of algorithms transport research, data training model to achieve the hardware, as well as positioning, processing, radar, laser radar and the like processing. Domestic team is responsible for defining the different functional requirements of each module, requires North American team to deliver on time and the quality of these features. In addition, poaching is also a response to the shortage of talent, a means of promoting the rapid development of technology, is also very common in the industry. However focuses on whether people still want to dig even people and “technology” are reluctant to let go? 2, the technical barriers in front of a lot of ink used to demonstrate excellence in the field of automatic pilot shortage, and lack of human resources is also a way nature means that a very high technical barriers, automatic driving skills is the case.
Fusion autopilot variety of techniques involving a plurality of fields, including image recognition system algorithm imaging, machine vision, visual perception, large data processing, object detection, etc., involved in the project development process may even multiple patented of thousands, all you want is actually very difficult to develop by its own. In addition to the basic patent, automatic driving technology itself is perceived by the surrounding environment sensor mounted on the car, then analyzed by the data processing environment surrounding the vehicle is equipped with “intelligent brain”, and so the driver is determined to make decision-making, control the direction of vehicle travel and so on. this processHow to them, the use of sensors distributed, the amount of data collected, the amount of data and other tests are very critical technical difficulties. And, because of the huge amount of data collected, the computation process will be more complex, “smart brain” process pressure will be enormous, while maintaining awareness of how the optimal combination of sensor to enhance the operation efficiency is The core difficulty. In the decision-making control, when the sensors to collect environmental data, “intelligent brain” to process the data, plan routes direction of travel, the route to be updated in real time with the car the environment, conditions and requirements of low latency, then for the amount of data processing required is very high, which is the biggest difference between the different autopilot companies. At the same time, collected large amounts of data adjustment is also very important to bring the results of different data inputs are completely different. There is an article on this is directed at skilled theft autopilot core reason: “When a technician training at a company a set of sophisticated autopilot algorithm, for after-service all the means to start over, which is a lot of people the reason to take the risk away of research results. “when we are eyeing the domestic aspects, technical barriers but also from policy reasons for this level. For example, to obtain the map of Surveying and Mapping qualification is extremely difficult, since the end of 2001 to 2018, China was only just 20 units received the “pass.” Also included autopilot drive test license release, the related art deflection requirements. This is also the reason prompting many domestic enterprises to set up R & D centers autopilot abroad.
Finally autopilot skilled frequent theft Another reason is that, behind the quick success of individual companies. When Tesla sued its Autopilot charge of violation of noncompete agreement poach staff team, bluntly in the trial, the industry has created investment into a quick profit, the entire industry has become impetuous and bubble. Profit-driven nature of the business has re-iterate CONTROL ENGINEERING China Copyright , and individual companies in order to obtain financing or to achieve rapid development in technology beyond, thus “to promote” technology theft, exaggerated publicity, in fact, the final result is difficult to escape the “bubble fragmentation” of. Having said that the reasons, which have a way to solve it? Avoid “theft” need the talent and innovation of the current autopilot is still in the preliminary stages of development, regardless of technology, products, markets and so onThe need to constantly improve and nurture. For autonomous vehicles, it can sometimes even now through technology “theft” to pay copyright fees to gain some advantage, but from the beginning of last year hot “chip” will be able to know, in the long run, competition in the market, only to master the product the core competence of technology to be able to have the last laugh, not to be various factors limiting the development of the company itself. Where the talents, the introduction of open learning and research cooperation is very important. For the field of personnel training autopilot, personnel training cycle is longer on the one hand, the other hand, companies need to be able to directly enter the company’s employees work takes a lot of time training rather than “white.” From this perspective, can make use of the resources of the media and universities, and guide students to apply the relevant professional at the same time as the field of artificial intelligence as in schools, additional relevant professional personnel to fill the gap. And promote the formation of cooperation between industries, enterprises and universities CONTROL ENGINEERING China Copyright , to achieve true integration of industry, in addition to the theory so that students can practice with real practice. Of course, for some large enterprises, gathered talents in the field of autonomous driving up overseas R & D center in Silicon Valley or will direct the gathering of such people “grab talent”, the introduction of talent is one way to resolve.
It also needs to be emphasized is that business for independent innovation, intellectual property value, which can be achieved through increased research and development efforts to investment and acquisitions, investment and other means. With BAIC, for example, it had announced that it would invest 2 billion yuan in 2020, invested 5 billion yuan by 2025 technology research and development; Xiaopeng car, Wei to automobile also said it will expand research and development team. From a policy perspective, the current domestic various provinces and cities have gradually liberalized the issuance of autopilot drive test license. Case Study of Beijing CONTROL ENGINEERING China Copyright , in which it issued last year a total of 54 automatic driving license test, other relevant policies has been steadily promote them. In addition to innovation, protection technology for businesses is also essential. Enterprises can Jingye agreement to form a relevant security team, instant patent application, such as for technology to enhance vigilance to prevent the theft of their own research and development of technology theft. To Baidu, for example, that once inAfter hackers found their own autonomous driving technology, set up a team of students from Tsinghua University, allowing students to help them find a loophole where their systems. Meanwhile, Apple’s technical staff after the discovery of the theft were also made to all employees a long warning memo , warned employees not to disclose the data to the media. The letter, Apple revealed only in 2017 that it arrested 29 leakers, in which 12 people were arrested and prosecuted. All in all, technology ten million, the first innovation, theft desirable, two lines of relatives of tears.