Automated Driving 2024: Navigating the Path to Mass Market Autonomous Mobility

Prof. Shai Shalev-Shwartz, CTO of Mobileye, delivered the keynote speech at Automated Driving 2024 in Frankfurt.
Keynote title: Navigating the Path to Mass Market Autonomous Mobility
Description: Propagating autonomous vehicle technology to a mass market demands an intricate strategy addressing both technological and regulatory obstacles. Critical technological impediments encompass matters of safety, scalability, and affordability. Regarding regulation, the significant hurdles involve defining an acceptable safety threshold and establishing a validation process for this definition. In this presentation, I will delineate our approach, placing emphasis on several fundamental principles: such as "modular design", "elimination of failures-by-design", "geographic scalability by crowd sourcing", "safety models", "redundant paths", and "crowd sourced validation".
0:00 Introduction
0:10 Mobileye SuperVision
3:20 Going from eyes-on to eyes-off
5:14 Challenges to reaching mass market autonomy
9:11 Solutions to build a scalable eyes-off system
13:36 Evaluating new technologies used to build AV systems
19:03 How to properly use AI for self-driving
20:12 Overview of redundancy
25:50 Problems associated with end-to-end perception
28:46 Bird’s-eye view (BEV) systems advantages and disadvantages
34:52 Summary
36:49 How does Mobileye tackle driving in challenging conditions?
40:54 Attaining high levels of SPI & independence of redundant systems
45:37 Increased development costs due to global variations in safety KPIs
About Mobileye:
Mobileye is leading the mobility revolution with its autonomous-driving and driver-assist technologies, harnessing world-renowned expertise in computer vision, machine learning, mapping, and data analysis.
Our technology enables self-driving vehicles and mobility solutions, powers industry-leading advanced driver-assistance systems, and delivers valuable intelligence to optimize mobility infrastructure. Mobileye pioneered such groundbreaking technologies as True Redundancy™ sensing, REM™ crowdsourced mapping, and Responsibility Sensitive Safety (RSS) technologies that are driving the ADAS and AV fields towards the future of mobility.
Connect with Mobileye:
WEBSITE: www.mobileye.com/
LINKEDIN: / mobileye
TWITTER: / mobileye
FACEBOOK: / mobileye
INSTAGRAM: / mobileye

[@z4zuse]

The good/bad data separation does not need to run real time. The glue code to tie together NNs does. On top of that real time glue code runs on CPUs, NNs run on inference hardware. Not a fair comparison.

[@rune6119]

He said even If it runs offline, its still gluecode. It seems your mental capacity was not sufficient to follow along.

[@z4zuse]

Given the readability of the slides this could have been a podcast

[@OutdoorEsports]

I think he has been giving the same speech for a decade

[@strategicweapons]

please hire better camera man for this . your power point is barely visible. If mobileye's camera system doing the job like this, there will be car crash everywhere.

[@netscrooge]

The PowerPoint slides could be edited directly into the video frame. No cameraman required. They were just too lazy to do it.

[@ranxerox10]

Seems like Mobileye is doomed, i see stagnation and no clear path to the system they want to create. much marketing but no real path.

[@marketsareopenmao]

In what sense?

[@tomcockcroft9394]

Yes, wish they would just show 100 hrs of the product like Tesla does. But all they have done in 5 years is two drives

[@speedismyfriend]

You don't seem to know much about Mobileye. They have the clearest path out of anyone in the consumer facing industry, with a step by step of how you advance from L2 to L4 and how OEMs can adopt it into their vehicles.

[@speedismyfriend]

@@tomcockcroft9394 They have 110,000 Zeekr vehicles on the road using their Supervision system. They have L4 robotaxis testing in multiple cities...

[@ranxerox10]

@@speedismyfriend they need data, i dont see how they get data