What technologies will Mobileye and Intel offer for autonomous driving applications? How will the two companies divide their work?
On a highly automated vehicle platform, the relationship between Mobileye and Intel seems to be getting closer and closer, which is worthy of caution for other competitors in the market or for those who intend to enter.
Evidence A is a strategic alliance between BMW and Mobileye and Intel in July this year. Evidence B is the recent auto parts manufacturer Delphi announced that it will use Intel chips and Mobileye's visual SoC on the autonomous driving platform; after all, Mobileye is already a leader in visual processing technology, a key component of advanced driver assistance systems (ADAS) and autonomous vehicles.
Intel has not announced that it will be equipped with multiple Xeon core automotive SoCs, but according to the rule of thumb, the world's largest processor supplier's skill can not be underestimated; in fact, Intel CEO Brian Krzanich before the US Thanksgiving Day In a keynote speech at the LA Auto Show, the company has invested $250 million with the goal of auto-driving vehicles.
In another article in the recent article, I discussed whether Intel can win the battle in the car chip war. Some readers’ intuitive response is that journalists have the conditions to believe and assume that Intel will succeed in the non-PC market. The most sensible opinion came from Philippe Lambinet, the former executive of any law semiconductor (ST) and currently CEO of Cogito Instruments.
Cogito Instruments is a manufacturer of machine learning analysis technology for industrial applications; Lambinet believes that the true value of Mobileye lies in datasets and algorithms: "They develop hardware because there is nothing to do the work they need, but basically hardware. Not their core value. On the other hand, Intel is a processor manufacturer."
Lambinet calls the combination of Mobileye and Intel "the perfect functional mix". He believes that Intel will establish a "dutel" situation like "Wintel" in the automotive market, trying to reproduce the company's past in the PC. The market has been brilliant with Microsoft. But Lambinet doesn't think Mobileye will have the same idea as Intel; it's a fun subject worth watching.
Indeed, when we carefully examine the technologies provided by Mobileye and Intel in the field of autonomous driving, what are the contributions of both parties? How will Mobileye's EyeQ chip and Intel's unpublished automotive SoC be divided?
To this end, I contacted Daniel Galves, Senior Communications and Senior Vice President of Mobileye. I asked the above questions on the Delphi platform. Galves' answer was: "It is very simple to start the union. Mobileye SoC will execute all sensor processing software. (8 camera surround video developed by Mobileye, radar / optical processing developed by Delphi), in addition to the mobileiza REM development of localizaTIon mapping (localizaTIon mapping), and sensor hub."
He went on to say: "Intel's SoC will implement all driving strategies (the path strategy enhanced learning algorithm developed by Mobileye) and driving control (driving behavior software developed by Telto's OttomaTIka)." This means EyeQ5 (expected in Will the engineering samples be available in 2018) Will the driving control software be implemented as well? Does EyeQ5 also need to match a chip like Intel is currently developing?
Galves said: "In the driving control software part, it is very computationally intensive, we did not play a very important role; therefore, this part is reasonable to operate on a separate chip, and we develop the relationship between path strategy and driving control. Very close, it works reasonably well on the same chip;" he also pointed out the benefits of operating on two different chips, including funcTIonal safety and redundancy.
One analyst speculated that the close relationship between Intel and Mobileye may threaten the current relationship between ST and Mobileye; Mobileye's Galves insists that it is impossible, he said that ST is an important partner for EyeQ4 and EyeQ5 to compete: "Our algorithm is built with the most efficient operation of the architecture developed by both parties. Delphi's OttomaTIka software is developed in the Intel environment, so deployment is reasonable in this type of architecture."
The bigger problem is in the future; beyond the Delphi platform, what are the legitimate expectations of the industry? What steps will Mobileye and Intel take to make a profit from the autonomous vehicle platform?
Lambinet is expected to have a profit war. His view is that Mobileye has a very good chance of achieving oligopoly in automotive applications: "So this is inevitably very attractive to Intel," with history as a mirror, which is why Intel and Mobileye are taking the lead. The closer the reason, but it is not clear whether Mobileye is so willing to go on like this.
"Intel became Intel today because of Wintel's business model. Microsoft and Intel once monopolized the PC market with a model that was very profitable for both parties. Since then, Intel has been trying to replicate this model. Because they can't reduce gross margins;" Lambinet said: "In the field of mobile phones and TVs, Intel has failed. The company may also fail again in the automotive field because the PC model is an exception rather than a rule."
Lambinet believes that if Mobileye's algorithm continues to dominate the autonomous vehicle platform, Intel is a hardware supplier: "If you don't buy Mobileye, you will be commoditized." But it's hard to imagine that Intel will not go through one. Struggling to take such a route.
Intel is now locking in artificial intelligence, using Nervana Systems to train neural networks (currently the field is dominated by Nvidia's graphics processors), and acquiring unfinished Movidius to handle computer vision and edge networks... we are once again in this field. I believe Intel will succeed?
Compilation: Judith Cheng
(Reference text: Is Mobileye-Intel New 'Wintel' of Auto?, by Junko Yoshida)
Rack-mounted Energy Storage Lithium Battery
Enershare's commitment to future-ready energy solutions for smart home innovations, Enershare`s Energy Storage Systems create a flexible energy maintenance system for homeowners who want to take more control of their home energy use, it is intended to be used for home battery energy storage and stores electricity for solar self-consumption, load shifting, backup power, and off-the-grid use. you can use it anytime you want-at night or during an outage.
Lithium Battery Storage,Lithium Energy Storage,Off Grid Solar Power,Lithium Energy Storage System
Shenzhen Enershare Technology Co.,Ltd , https://www.enersharepower.com