MobilEye announces plans for robotaxis in Munich

Topic: 
Tags: 

Companies are at the stage of announcing real pilot projects for robotaxi service. Now MobilEye announces they will start a robotaxi service in Munich and Tel Aviv by 2022. What are the new metrics of success for a team?

See more at MobilEye announces plans for robotaxis in Munich

Comments

deepai dot org/publication/safe-perception-a-hierarchical-monitor-approach

Perception systems of automated vehicles rely on AI-based detection algorithms. In general, these algorithms offer great detection quality, nevertheless they rely on the quality of the training data and thus there is always the possibility that objects are missed even though they are clearly visible, simply as these were not covered in any training data.

Consequently, it is hard to construct a proof that such an AI-based system is safe under all operating conditions. To address that gap, we presented a novel monitor to check the correctness of a given perception system. The monitor uses LiDAR information, and is realized as a combination of a rule-based probability filter, that allows the system to correctly identify prominent objects that may have been missed by the primary perception system, and additional filters that both increase detection rate and decrease false alarm rate.

We highlighted in our evaluation that a state-of-art AI detector for LiDAR perception can miss prominent and safety relevant objects in the surrounding of the vehicle. Also these detectors are sensitive to configuration parameters. A small change in the configuration of the confidence threshold can lead to miss detection of clearly visible objects. The use of lower confidence can reduce the miss detections, but only at the cost of more false alarms.

With our monitor architecture we are able to eliminate the majority of detection misses, and the few remaining ones are all heavy occluded and not in direct connection to the ego vehicle (thus not safety relevant). In addition to that the false alarm rate is still low ( <2%).

Hence, in summary, with our approach we can improve safety of the LiDAR perception system and still maintain a high availability. It is worth mentioning, that in this work we focused on a single channel LiDAR system. In a final AV, it is desirable to use additional modalities in parallel, to gain robustness against sensor failures.
Intel Labs Europe

Mobileye has about 3,100 employees.

In its filing, Mobileye noted that it acquired mobility and transportation business Moovit from Intel this year.

Mobileye has collected data from 8.6 billion miles

Mobileye has shipped 117 million EyeQ.

As of July, it had $774 million of cash and cash equivalents. In the 12 months ended Dec. 25, it had a net loss of $75 million on revenue of $1.39 billion.

Mobileye said in its filing that its board will include Gelsinger as chairman, and also former US Senator Claire McCaskill, a Missouri Democrat, and Jon Huntsman, the former Republican governor of Utah as well as ambassador to China who is now on Ford Motor Co.’s board.

Mobileye also plans to have four Intel-affiliated members on its board, including Chief Executive Pat Gelsinger serving as chairman of Mobileye’s board.

Intel will also get paid from the offering: Mobileye issued Intel a dividend note for $3.5 billion, and expects to pay that off with proceeds from the sale, according to the filing; there was an initial payment of $336 million, leaving more than $3 billion still owed to Intel.

By 2030, the company expects its driver-assistance systems to be deployed in another 266 million vehicles globally.

(if 70 million cars sold annually, Mobileye expects 45%-48% cumulative marketshare up to 2030 ?)

The company estimate the current total addressable market (“TAM”) to be approximately $16 billion, composed entirely of selected ADAS market opportunities. They expect the near-term TAM to be approximately $40 billion and the long-term TAM to be approximately $480 billion, as the value of ADAS functionality increases and as Autonomous Vehicle (“AV”) deployment, both in consumer-owned vehicles and fleet-owned vehicle networks, accelerates.

After Mobileye’s public debut, Intel will beneficially own all of the outstanding shares of Mobileye’s Class B stock. “As a result, Intel will be able to control all matters submitted to our stockholders for approval, including the election of our directors and the approval of significant corporate transaction,” the filing stated.

In the regulatory filing, Mobileye said that it had approximately 3,100 employees operating across eight countries, with approximately 80 percent of its workers in research and development and approximately 2,900 of its employees operating in Israel.

www dot sec dot gov/Archives/edgar/data/1910139/000110465922104640/tm227410-15_s1 dot htm

lots of "new" info scattered throughout (actually requires reading much of the document to pick out all the new details and nuances.

IE. new details NOT confined to (the normal) Prospectus Summary, Risk Management, and Management Discussion sections.

New details are spread throughout and helps to read all.

sidenotes
"... driven by regulatory requirements and safety ratings of a Mobileye SuperVision-like solution in its own category. " ???

- "... New central compute ECUs with Intel ..." in the works

- cancelled project in June 2021 with Transdev

2021 Ford BlueCruise uses Mobileye

S1 and website
Over 14x growth in 30 months?

REM enabled vehicles harvesting goal is to grow from 1.5 mil to 21.67 mil in 30 months (over 20 million in 30 months or 695,000 per month, or 2.1 million per quarter)

or in other terms,
227 billion miles harvested per year by beginning of 2025 which equals 622 million miles per day.

Cumalative miles by July 2022 was 8.6 billion REM harvested miles and a run rate of 43 million per day.

Goal of over 14x growth

A minimum of 2 million REM enabled (cloud enhanced) vehicle sales per quarter is roughly or minimally about 20 to 25% of company sales.

All eyes on Q3 earnings to see if REM data is updated now that Mobileye S1 is out.

The prediction of 21.67 mil vehicles per day that will harvest REM data by the end of 2024 is still on the website.
This goal has not changed since CES 2021 presentation.

source for
vehicles / miles harvested per day
CES 2021 conflicting / 5 mil per day
CES 2022 1 mil REM / 25 mil per day
S1 2022 1.5 mil REM / 43 mil per day
Q4'24 21.5 mil REM / 622 mil per day

Cumalative Jan 2021 - 4.66 bil miles
Cumalative Jul 2022 - 8.6 bil miles
Cumalative Jan 2025 - 227 bil miles

CES 2022 shows about 1 million vehicles harvesting in Europe in 2021 using the assumption of 28.67 miles per day.

Definitely exponential from 0 to 227 billion in 7 years.

227 billion = 622 million miles per day multiplied by 365 days.

28.67 miles per day x 21.67 million vehicles

Best question: does a REM enabled vehicle have to be actively harvesting (subscription active) to be considered in REM statistics?

Goal 227 billion miles cumalative by Jan 2025.

there is an error in the assumption of the miles driven per day as this metric varies from country to country.
The target size of the fleet to generate 1 billion harvested KMs per day cannot readily be determined without a baseline average of mileage per day per vehicle based on some global average.
Where did 10,400 miles per year originate from?

The Mobileye goal is more likely to average adding about 500,000 vehicles to the harvesting cloud every month all the way through 2025, which is probably about 15% of the units sold through that entire 42 month period (if current goal is still harvesting 1 Billion KMs per day by end of 2025).

But Mobileye predicted in Nov 2019 to have 1 million vehicles harvesting in Europe by end of 2020, and 1 million vehicles harvesting in the US by end of 2021. But only 1.5 mil vehicles were collecting by Jul 2022 (did 500,000 vehicles stop or never start collecting).

REM harvesting data
(million veh / million miles per day)
CES 2021: ~1 mil veh / 5 mil per day
CES 2022: 1+ mil veh / 25 mil per day
Ju 2022: 1.5 mil veh / 43 mil per day

Per CES 2022, in all of 2021, 2.5 bil miles collected (cumulative) in Europe.

At CES 2022, The access to 1.2B KM mapping data improves (per VW HDiess)

REM stats
Cumalative Jan 2021 - 4.66 bil miles
Cumalative Jul 2022 - 8.6 bil miles

The official press release may have an error from CES 2021:
- "and nearly 1 billion kilometers completed to date"
- where as others reported cumalative of 7.5 billion KMs (4.66 bil miles).

Professional reporter below -
another source for CES 2021:
"More than 1 million vehicles have now built up more than 700 million km (435 million miles) of physical maps with 8 million more KM (5 million miles) being "updated" daily".

Current unique Jul 2022 from S1:
"We estimate that the data we have accumulated covers over 90% and 80% of the approximately 0.8 million miles of motorway, trunk, and primary road types in each of the United States and Europe, respectively."

??? Not sure on the above from S1.

Some details provided well in advance in the YT videos.

Tuesday, Nov. 5, 2019
1) Autonomous Vehicle Sensing Tech and Algorithms with Gaby Hayon of Mobileye

Gaby Hayon, executive vice president of research and development at Mobileye, speaks Tuesday, Nov. 5, 2019, at the 2019 Mobileye Investor Summit at its headquarters in Jerusalem.

2) Mapping Technology for AVs with Tal Babaioff of Mobileye

Tal Babaioff, Vice President of Mapping and Localization and Co-General Manager REM at Mobileye, speaks Tuesday, Nov. 5, 2019, at the 2019 Mobileye Investor Summit. Mobileye, at its headquarters in Jerusalem

Intel-Mobileye is expected to detail its new valuation expectations Tuesday.

"Mobileye is now set to target one that is under $20 billion."

Mobileye plans to launch its roadshow for prospective investors on Tuesday.

The goal is still for the shares to begin trading Oct. 26.

www dot wsj dot com/articles/intel-eyes-significantly-lower-valuation-in-ipo-of-mobileye-unit-11666037700?

Mobileye Global Inc., expects a market valuation of roughly $15 billion at the projected pricing midpoint of its initial public offering.

In a prospectus filed with regulators, Mobileye said it plans to offer 41 million class A shares at a projected price of $18 to $21 apiece.

At the expected $19 a share midpoint of the IPO's pricing, Mobileye said it expects net proceeds from the offering of about $800 million, or $900 million if the underwriters' option is exercised in full.

Many are baffled why they even want to go out at such a valuation, similar to the valuation when Intel bought them. Why not wait for the market to get out of bearland? I guess this will let them give more lucrative stock options to recruit.

Employee compensation in Mobileye shares was one of the factors that pushed Mobileye founder Amnon Shashua to go public. This is in view of the ongoing drop in Intel shares, which lost more than half of their value in the last year and made it difficult for the company to attract engineers.

All the shares in the offering will be sold by Intel, which will own 94.2% of the company after the completion of the IPO. This is the sale of a very small share of Mobileye - less than 6% - and Wall Street investors historically do not like IPOs that offer so few "goods". It seems that the meager number of shares is the result of a compromise between Intel, which would have preferred to wait a little longer for better days, and Shashua, who is pushing for an IPO to strengthen the brand of the company he founded to make it easier to deal with the intensifying competition from strong rivals such as Qualcomm and Nvidia. Both competitors are public companies that publish not only their results, but also technological announcements, big and small, and Mobileye feels that its activities are overshadowed by Intel.

www dot calcalistech dot com/ctechnews/article/s1bxkvamj

Shashua's desire to "be released" from the Intel prison and turn the spotlight on Mobileye's technology can be read between the lines of the letter he sent yesterday to 3,100 Mobileye employees. "Today we are starting the road show for Mobileye's IPO. The move will complete almost a year of preparations for Mobileye to become a public company. A year in which we saw tremendous progress in our performance and product portfolio, with an emphasis on Robotaxis. As a public company, we will be subject to market fluctuations, however, more importantly, we will receive credit for the unique value we are creating - the value of the technology, of the strategic plans, of the execution and of the financial growth," Shashua wrote. He then added an example. "To put things in context, imagine that the IPO is the first 100 meters in a marathon. What is important is where we will be many kilometers after the start of the race and not necessarily at the starting sprint. We are launching the IPO in very challenging times for global markets, but we believe we need to focus on the horizon. That's where our growth is."

Intel will wait for a recovery in the markets and an increase in the stock to sell another package of shares without an additional prospectus in about six months.

there is additional analysis in article

www dot calcalistech dot com/ctechnews/article/s1bxkvamj

how come u do not repeat last third of article

no mobileye robotaxi in israel or germany so far and 2022 almost over

In July 2022, Mr. Dagan resigned from his position as Executive Vice President, Products and Strategy to take on an advisory role and continues to be an employee of the Company for a temporary period of transition of his responsibilities. He is no longer deemed to be an executive officer of the Company.

increase.

The difference between Mr. Dagan’s 2020 salary and his 2021 salary is the result of unpaid absence days taken in 2021.

When the EVP of Products and Strategy resigns, there is much the new shareholders of Mobileye will never know.

And now Shashua is starting another company, Mentee Robotics, per S1 reg.
Chairman and co-founder of Mentee Robotics.

Where does one find a new EVP for Products and Strategy for a company such as Mobileye ?

will new EVP come from Auto OEM, Tier 1, Tech or Intel?

Erez Dagan was responsible for Mobileye’s business objectives, the company’s product planning and architecture, strategic planning and initiatives, business development, and advanced development projects. Under his leadership, the Product and Strategy division manages Mobileye ADAS & AD partnerships with OEM customers, as well as Mobileye's Self-Driving-System program management and partnerships with driverless MaaS customers. Dagan is also responsible for Intel's LiDAR and imaging radar business units as well as the LiDAR-Radar algo sensor fusion department.

why did intel-mobileye spend 900 million on moovit

before the AV finally launches?

Yet - still listed as "Executive Vice President, Products and Strategy, Mobileye
Vice President, Intel Corporation" on Mobileye's website of mgmt

With promotions of
Johann Jungwirth in July 2022 to
Senior VP, AV
and
Nimrod Nehushtan in July 2022 to
Senior VP, Strategy / Business Devel
and
Yossi Hadad in July 2022 to
Senior VP, SW/HW Integration/Release
and
Boaz Ouriel in April 2022 to
Senior VP, EPG Software
and ...

Yossi Kreinin - Oct 1
"Bad " isn't the problem - at least it's never your only problem by the time you realize you have a problem. At that point, you have a bad product made in bad ways by a badly structured organization. Throwing brilliant people into this shit will not turn it into honey"

Aug 10
"You know that sinking feeling when meeting someone smart, energetic, knowledgeable, opinionated, megalomaniac and having no common sense whatsoever? About to break that which nobody has tried breaking yet, and to build that which nobody will be able to deal with?

Ohhh shiiit."

www dot sec dot gov/Archives/edgar/data/1910139/000110465922104640/tm227410-15_s1 dot htm

only 20 mil design wins per yr for 8 yrs ???

266 m - 87 m = 179 mil ÷ 8yrs = 20m

We estimate, based on our existing design wins through July 2, 2022, that our ADAS solutions will be deployed in more than an additional 266 million vehicles by 2030, including approximately 37 million vehicles based on our first half 2022 design wins and approximately 50 million vehicles based on our 2021 design wins.

"Consumer AV competitors include Apple, Sony, and Tesla, who are developing self-driving vehicles for consumers."

The semiconductor industry is experiencing widespread shortages of substrates and other components and available foundry manufacturing capacity, and we anticipate that such shortages will continue. These factors, combined with the long lead times associated with wafer production, have contributed to a shortage of semiconductors. During 2021 and through the first half of 2022, STMicroelectronics N.V. (“STMicroelectronics”), our sole supplier of EyeQ SoCs, was not able to meet our demand for EyeQ SoCs, causing a significant reduction in our inventory level, and we expect to continue to experience a shortfall of chips during the second half of 2022. We have entered 2022 with significantly lower inventories of our EyeQ SoCs as a result of the limited supply during 2021, and, due to continuing supply chain constraints, we are operating with minimal or no inventory of EyeQ SoCs on hand. As a result, we are substantially reliant on timely shipments of EyeQ SoCs from STMicroelectronics to fulfill customer orders and are unable to offset future supply constraints through the use of inventory on hand, In addition, without a solution to the shortages, we may continue to have insufficient inventory in subsequent fiscal years. Since our EyeQ SoC is the core of our ADAS and autonomous driving solutions, continued shortages in the supply of sufficient EyeQ SoCs to meet our production needs would impair our ability to meet our customers’ requirements in a timely manner, and would adversely affect our business, results of operations, and financial condition. The limited supply of EyeQ SoCs has already led to rescheduling deliveries to our customers on certain occasions and may continue to cause delays in our ability to fulfill our customers’ orders as scheduled

Simple. Mobileye is providing their pipeline assessment through July 2,2022.

"These estimates may DO NOT include design wins AFTER July 2, 2022."

"We estimate, based on our existing design wins THROUGH July 2, 2022, that our ADAS solutions will be deployed in more than an additional 266 million vehicles by 2030, including approximately 37 million vehicles based on our first half 2022 design wins and approximately 50 million vehicles based on our 2021 design wins."

The 266 million unit pipeline may be pipeline is already 36% of ADAS marketshare up through 2030.

Based on 740m light vehicles sales in next 8 1/2 years.

Shashua publically stated he does not believe in L3 but Mobileye has designed Chauffeur as a consumer AV solution that may require driver intervention in certain situations. That is a contradiction.

40 new design wins in 2021 estimated to generate 50 million vehicles to be shipped by 2030. And an unknown number of design wins in 2022 should produce 37 million vehicles shipped by 2030

The pipeline differential of 115 million from below illustrates unless the cumulative pipeline is provided (266m as in S1 reg), the value involves effort to establish.

2017 Pipeline volume of new design wins = ?? million (30 new design w)
70 vehicle models launched

2018 Pipeline volume of new design wins = ?? million (28 new design w)
78 vehicle models launched

2019 Pipeline volume of new design wins = 28 million (33 new design w)
16 vehicle models launched

2020 Pipeline volume of new design wins = 36.2 million (37 new design w)
??? vehicles launched

2021 Pipeline volume of new design wins = 50 million (40 new design w)
188 vehicles launched

H1'22 Pipeline volume of new design wins = 37 million (? ne
w design w)
??? vehicles launched

"AV competitors include Apple, Sony, and Tesla, who are developing self-driving vehicles for consumers."

Singling out Apple and Sony as AV competitors is unusual.

Could Sony teaming up with Honda be more substantial than most consider (known as Sony Honda Mobility Inc)? Does Mobileye know something (eg threat from new Sony intelligent sensors)?
See Autonews article on new JV.

On a funny note, from a auto blogger who is a hoot
"Sony EV might end up being built by Honda, but on a GM platform."

But seriously, Sony is going all in on intelligent sensors.

Sony plans to spend about $7 billion to develop imaging sensors in the 2021-23 period. And Sony expects to supply imaging sensors to 15 of the world’s top 20 global automakers by 2025.

Sony Group is developing a sensor for self-driving vehicles that uses 70% less electricity, according to reporting from Nikkei Asia.

The sensor will be made by Sony Semiconductor Solution and will be paired with software developed by Tier IV. The companies aim to deliver Level 4 autonomous tech by 2030.

Sony plans to lower the amount of electricity needed for the sensors by relying on edge computing. The company plans to process as much data as possible through AI-equipped sensors and software on the vehicle itself, instead of transmitting that information to an external network.

Sony said this approach should make autonomous vehicles safer, as it will shrink communication lags. The company also plans to incorporate image recognition and radar technologies into its sensor to make a self-driving vehicle better equipped for rain, fog or other adverse conditions.

Autonomous vehicles use large amounts of power because of all of the added technology on board, which can result in at least a 35% smaller range for electric vehicles (EVs), according to Japan’s Ministry of Economy, Trade and Industry. If Sony reaches its goals, it could limit this impact to just around 10%.

The company’s goal is to reduce the amount of power used by an electric vehicle’s on board systems by 70%.

Sony also is also improving high resolution sensors. “If I want to see 500 meters away, then I can use a specific small part of the image, kind of like a virtual zoom.” He explained that detecting the color of a traffic light several blocks away could help determine broader traffic patterns and thereby suggest if a car should slow down or maintain its pace.

NOTE from Mobileye S1 reg: "Additionally, we intend to explore a collaboration with Intel on a technology platform to integrate our EyeQ SoC with Intel’s market leading central compute capability, with plans to utilize Intel Foundry Services’ advanced packaging capabilities. This potential platform is intended to enable functions essential to safety, entertainment, and cloud connectivity".

"intend to explore" means little.

Considerable new details about Mobileye scattered throughout the filing. And for a definition of the new SuperVision Lite, requires aggregation pulled together from various sections in the document.

Note also:
Additionally, we intend to explore a collaboration with Intel on a technology platform to integrate our EyeQ SoC with Intel’s market leading central compute capability, with plans to utilize Intel Foundry Services’ advanced packaging capabilities. This potential platform is intended to enable functions essential to safety, entertainment, and cloud connectivity.

However, we believe that the expected continued constraint on global automotive production resulting from supply chain shortages and the effects of economic uncertainty will limit our ability to increase our revenue.

..
We will be adding a new innovative Premium ADAS Solution, SuperVision Lite.

....
The complete definition of SuperVision Lite is spread out in 3 seperate sections of the document and needs a joined composition to clearly flesh out the details and scope.

And specifics of ALL of the 4 below spread out in 3 seperate sections of the document and needs to be aggregated per each solution to clearly flesh out the details and scope of each approach.

Mobileye SuperVision Lite
Mobileye SuperVision
Mobileye Chauffeur
Mobileye Drive

The motivation of the IPO this year is getting cloudy.

Company admits revenue growth an issue for 18 months (through 2023).

Item 1 - From S1 - "However, we believe that the expected continued constraint on global automotive production resulting from supply chain shortages and the effects of economic uncertainty will limit our ability to increase our revenue."

Item 2 - next, read the post above titled "minimal or no inventory"

Item 3 - ASPs going up as cloud based ADAS sales increase to offset cap on EyeQ sales ( $200 million bump to ASP from cloud based assist and $200 million from organic growth (max out production). 2022 may finish capped around 33 million total EyeQ sales, and $1.85 billion in total revenue. Company could have sold 36-38 million EyeQs and reached $2 billion in sales had STM been able to deliver (as Mobileye has commented that inventory on hand over the previous years has averaged 10-15% which currently is non-existent per S1).

The REM enabled fleet growth from 1 million in Jan 2022 (CES) to 1.5 mil in Jul 2022 (S1) is "curious" since most of the European uptake began Q3 of 2021 (see table below). Given Europe 2021 stats (see slides Jan CES 2022 and S1 (July 2, 2022 states 1.5 million REM-enabled vehicles), the new 500,000 in 2022 may be weighted towards Europe. Regardless, 500,000 additional units could be as high as $200 million which offsets lost revenue from supply issues. At least historical data is available given Mobileye REM predictions failed from CES 2020:
Over 1M Harvesting veh in EU by 2020
Over 1M Harvesting veh in US by 2021

Mobileye's REM/Roadbook presentations over the last 3 years seem to be intentionally incomplete which is strange given REM is a fundamental pillar. Most would not realize 1 billion daily KMs collected is simply about 21,700,000 cars on road.

Base ADAS = $2-digit low
Base ADAS +cloud ADAS = $2-digit high
Enhanced ADAS = $3-digit low
SuperVision ADAS = $4-digit low
NEW SuperVision Lite ADAS = ????

Europe, harvesting (in miles) in
Q1 2021 = 153.5 million / 60,000 vehi
Q2 2021 = 292 million / 113,000 vehi
Q3 2021 = 957.5 million / 370,000 veh
Q4 2021 = 1087.4 million / 420,000 ve

Because Mobileye does not identify uniqueness in stats, one cannot assume 2021 European REM enabled vehicles are 970,000 (sum of year) vs 420,000 (run rate at end of year). The distinction cannot be determined by the information provided so wether there is 420,000 or 970,000 new vehicles harvesting data by the end of 2021 is not clear.

Item 4 - compare CES 2022 presentation slides and video combined together to the S1-reg. If the FMCW lidar is in "early stage" development according to S1 but the CES states late 2024 delivery, then EyeC radar could easily be available late 2023 (with SuperVision Lite?) to early 2024 as neither state in early stage development. Either way, EyeC radar is trending to be available a year ahead of EyeC FMCW lidar so far.

Change 1 should read

Regardless, 500,000 additional units could be as high as $400 million additional dollars which offsets lost revenue from supply issues, for a balance of $200 million additional dollars.

Change 2 should read
(as Mobileye has commented that the surplus of inventory on hand over the previous years has averaged 10-15% which currently is non-existent per S1).

net loss / revenue / EyeQ
2021 -$75 mil / $1.386 bil / 28.1 m
2020 -$196 mil / $967 mil / 19.7 m
2019 -$328 mil / $879 mil / 17.5 m

For the 1st six months of 2022
- $67 million / $854 mil / 15.9 m

For the 1st six months of 2021
+$4 mil / 754 mil / 14.4 mil

For the 2nd six months of 2021
-$79 mil / 632 mil / 13.7 mil

Research and development remains a large expense: Of the company’s 3,100 global employees, roughly 80% are dedicated to R&D. For the six months ending in July, R&D costs were $359 million, or around 42% of revenue. Typically, large chip companies spend about 20% of sales on R&D.

correction: 14.4 to 13.7 to 15.9

"we are developing Mobileye Chauffeur, our consumer AV solution with a human driver still in the driver’s seat that may require driver intervention in certain situations, and Mobileye Drive, our Level 4 autonomous driving solution."

Shashua blog on Medium:
difference between Level 3 and Levels 4–5
Since in Level 3 the take-over by the human driver is not instantaneous, it follows that from the perspective of required emergency maneuvers, Levels 3–5 are all the same. In other words, the first observation that we make is that from a system design perspective, Level 3 should handle immediate response of the vehicle autonomously at the same performance level as L4–5. This is true because a take-over request that gives the human driver a few seconds to react does not actually simplify the system design, because during those few seconds anything can happen.

amnon-shashua dot medium dot com/on-black-swans-failures-by-design-and-safety-of-automated-driving-systems-1401076e9027

This implies surround sensors will become a new standard of evaluation as Mobileye has noted in the S1.

Add new comment