Intel Purchases Autonomous Car Tech Firm Mobileye

[cageymaru]

Intel has purchased autonomous car technology company Mobilieye for $15.3 billion. Mobileye technology such as advanced driver assistance and anti collision systems are found in 70% of the self driving car global market. The two companies have been collaborating on developing a fleet of BMW autonomous cars that are scheduled to be road tested the second half of 2017. Mobileye has been purchasing it's chips from STMicroelectronics, and is currently assisting Intel with the design of it's 5th generation of autonomous driving chips that are scheduled for release in 2021.

Mobileye was founded in 1999 and received $130 million in investment funding from Goldman Sachs in 2007. Mobileye had an adjusted net income of $173 million in 2016 and was valued at $10.6 billion in 2014. This deal may be in response to the $47 billion Qualcomm purchase of sensor manufacturer NXP; which put tremendous pressure on the other players in the field.

Last October, Qualcomm announced a $47 billion deal to acquire NXP, the largest automotive chip supplier, putting pressure on other chipmakers seeking to make inroads into the market for autonomous driving components, including Intel, Mobileye and rival NVIDIA NVDIA.O. The Qualcomm-NXP deal, which will create the industry's largest portfolio of sensors, networking and other elements vital to autonomous driving, is expected to close later in 2017, subject to regulatory and shareholder approvals.

 

[kju1]

Telsa, now with Intel Inside.

 

[ICE_9]

/facepalm

This had better work.

 

[Deleted member 93354]

Telsa, now with Intel Inside.

Didn't Tesla drop mobile eye because they weren't innovating fast enough? Wasn't it Mobile Eye who released a statement after getting dropped that as a company "They were being pushed harder than what they felt as 'safe'"

From what I have been reading vector analysis of images is the fastest way to handle autonomous driving. I hate to tell you intel, but when it comes to SIMD, MIMD, SVX, etc ops...NVIDIA has you beat hands down.

 

[dgz]

Didn't Tesla drop mobile eye because they weren't innovating fast enough? Wasn't it Mobile Eye who released a statement after getting dropped that as a company "They were being pushed harder than what they felt as 'safe'"

From what I have been reading vector analysis of images is the fastest way to handle autonomous driving. I hate to tell you intel, but when it comes to SIMD, MIMD, SVX, etc ops...NVIDIA has you beat hands down.

Fastest in what way? I not nearly as versed as you but I am just wondering why are they putting all other kinds of active detection systems such as radar and LADAR? Are we at the point where they combine all this data into images?

 

[Deleted member 93354]

Fastest in what way? I not nearly as versed as you but I am just wondering why are they putting all other kinds of active detection systems such as radar and LADAR? Are we at the point where they combine all this data into images?

Well there's signal analysis (sensors), and vector analysis (movement), and object discrimination (Image recognition AI). All this information is integrated together. All of these use linear algebra which is a strong point of GPUs through the use of vector processors.

The more vector extensions you have, the faster a matrix breaks down. This means you can use more variables (more detailed analysis), or analyze a situation more often (reduced lag)

BTW: To answer your question: Chevy's uses stereo scopic image analysis ALONE. They discovered it is a ton cheaper than using a bunch of expensive radars.

 

[dgz]

Well there's signal analysis (sensors), and vector analysis (movement), and object discrimination (Image recognition AI). All this information is integrated together. All of these use linear algebra which is a strong point of GPUs through the use of vector processors.

The more vector extensions you have, the faster a matrix breaks down. This means you can use more variables (more detailed analysis), or analyze a situation more often (reduced lag)

BTW: To answer your question: Chevy's uses stereo scopic image analysis ALONE. They discovered it is a ton cheaper than using a bunch of expensive radars.

Thanks. Chevrolet's approach makes sense. After all that is how we manage. In fact our brains are really good at that. Personally, don't really like the idea of being blasted by microwaves on the street.