The information age is changing rapidly. Auto-driving cars have become a "terminal" market for technology companies, traditional car companies, and component manufacturers. It seems that any company claims to enter this field is not surprising.
One of the keys to the development of autonomous driving technology is the continuous improvement of the "autopilot chip" technology. At this stage, high cost, high power consumption, and insufficient computing power are the challenges that various types of vehicle-mounted "auto-driving chips" manufacturers are trying to solve. However, with the increasing upgrade of chip technology, the "automation" level of self-driving cars has jumped sharply, and the real "autopilot" may not be far away.
According to SAE InternaTIonal's classification system for autonomous driving, the automatic driving technology can be divided into 0 levels, 1 level, 2 levels, 3 levels, 4 levels, 5 levels, and a total of six levels. According to this grading standard, there is a considerable span between level 2 and level 3. When driving a vehicle of the first and second level of technology, the driver must always pay attention to the road conditions and respond in time. Nowadays, the functions of automatic anti-collision, cruise control, and automatic parking are all below level 2. The Autopilot assisted driving technology that Tesla is selling belongs to the second level technology. In the Level 3 technical standard defined by SAE, the task of monitoring road conditions is done by an automated driving system. Technicians also often view these two levels as the dividing line between "assisted driving" and "automatic driving."
It should be emphasized that even if the third level of automatic driving is implemented in accordance with the SAE standard, the driver must be on standby, ready to respond to system requests and handle special situations that the system is not capable of coping with. That is to say, when the vehicle reaches the L3 level of automatic driving, the human driver does not need to "drive", but he can't do anything else, still be vigilant, always ready to save in special circumstances. field". When the vehicle reaches the L4 level of automatic driving, all daily driving behavior is given to the vehicle's automatic driving system to complete, and the human driver can almost completely without supervision. This is the biggest difference between L3 and L4 autopilot.
However, whether it is L3 or L4, the requirements for "autopilot chips" are obviously higher. Therefore, chip technology has become the focus of the company that strives to show its talents in the field of autonomous vehicles.
At the recent 2018CES, Nvidia released the AI ​​supercomputing chip DRIVE Xavier. According to reports, it can perform 30 trillion deep learning calculations per second, and consumes only 30 watts. NVIDIA CEO Huang Renxun mentioned that a Xavier chip can provide undelayed computing support for L3-L4 autopilot vehicles. In addition, in 2017, international technology giants Intel, Qualcomm, AMD, Infineon, Google, domestic "Horizon", NavInfo and other companies have made more investment in this area. Delphi and MobilEye claim that they can provide autopilot systems that meet SAE Level 4 requirements by 2019.
In fact, some car companies have released mass-produced L3-class self-driving cars. The new Audi A8, which went on sale in 2017, can achieve L3 autopilot. However, because the L3 self-driving car still needs the driver to pay attention to the vehicle condition, Ford, Volvo and other car companies said that they will go directly to the L4 level beyond the L3 level.
These trends indicate that L3, L4-class self-driving cars will appear in large numbers in the near future.
Guess that in the next 3 to 5 years, the L5 self-driving car is expected to arrive.In fact, many professionals believe that from the perspective of chip technology, it is very difficult to cross from L3 to L4. Cheng Yabing, architect of Nvidia Solutions, believes that from L3 to L4 there is a 50-fold increase in the amount of computation and data, which is difficult to achieve based on existing decision-making algorithms and learning techniques. Obviously, to achieve this level of leapfrogging, technology companies and car companies need to invest more human, material and financial resources.
Although no company has publicly made relevant predictions on the technical difficulty of L4 to L5 chips, it is foreseeable that with the advancement of technology, the competition in the field of automatic driving will become more intense. At that time, for the L5 level automatic Driving technology will be more important, and the market will be laid out in advance.
On the hardware side, Nvidia has previously announced that it will work with AURORA, founded by former unmanned car executive Chris Urmson, to create a new L4 and L5-class autopilot hardware platform. In addition, Intel's Mobileye also announced that it will mass produce the fifth-generation ADAS (Advanced Driver Assistance System) dedicated chip EyeQ5 in 2018. According to reports, EyeQ5's computing power will be able to meet the L4-L5 level of autonomous driving. Tesla's founder Elon Musk also announced that the current Tesla car has been equipped with the ability to achieve SAE Level 5 autopilot on the hardware standard. He predicted that in 2018, Tesla will be able to provide electric vehicles with fully automatic driving functions, but it may take another one to three years before the model can be officially approved and marketed.
Based on the current information, the author believes that once the technology of the autopilot chip is qualitatively improved, the chip algorithm and speed are upgraded again after the L4 standard, and the relevant laws and regulations such as road rights and map information are improved, L5 level. Autopilot will be expected to be implemented as soon as possible, and this time will not be too long, and good news will come in 3 to 5 years.
Guess the three traditional car companies and technology companies against each other, which will be better?Today, technology companies, component companies, traditional car companies and many startups are entering the field of autonomous driving. In this case, how will traditional car companies and other types of companies get along?
Gong Yueqiong, general manager of Beiqi Foton Motor Co., Ltd. once said that traditional car manufacturers are hardware-based, mainly relying on R&D teams composed of thermodynamics and materials science, but in the era of artificial intelligence and electrification, car manufacturers need to have electrification and artificial intelligence. Data capabilities, which is more difficult for traditional vendors. The boundaries of the automotive industry have expanded from the original vehicle manufacturing enterprises to Internet companies and software companies. In terms of artificial intelligence of vehicles, the ability of Internet and component companies to be at the perception and execution levels is more than that of vehicle manufacturers. He stressed that "the automotive industry will be reconstructed in the future, and car companies must grasp the core of new energy and intelligence."
However, technology companies' new technologies, such as chip technology, need to be combined with the vehicle as a vehicle to be truly applied. For technology companies and Internet companies, they must also cooperate with car companies. For example, NVIDIA has already cooperated with Tesla, Volvo and other car companies; Intel also cooperates with BMW to study autonomous driving technology.
The author believes that in the future, in the field of automatic driving, the cooperation between traditional car companies and new technology companies will be far greater than the competition. As the saying goes, there is competition to develop and cooperation to develop faster.
LED Video Wall And LED Display Screens
We strive to create products which are environmentally balanced, outdoor LED walls are low power consumption.
Our outdoor LED displays are compatible with various input formats, such as DVD players, cable TV, internet & intranet etc.
These are highly customized LED display video walls, every LED display screen is customized regards to pixel pitches, resolution, sizes, shapes etc.Our LED display enjoys long life, they are weatherproof units & can tolerate dust, humidity or rain.
Due to DIP technology that is used the outdoor video walls are sturdy, durable & very bright. Unlike other LED video walls, Priva Outdoor Led Display are
viewable in direct sunlight. The picture quality is not compromised in
our solution, hence the display is crystal clear despite it`s a day time
or night time.
Priva LED display screen wall for outdoor applications is a module based LED video wall solution. The technology used for outdoor LED display is DIP. As this is a module based solution, it can be molded into any shape or size.
Led Display , Led screen, Led Display Screens, Led Display fixed, Led screen fixed, LED wall
Shenzhen Priva Tech Co., Ltd. , https://www.privaled.com