Check out Prof. Amnon Shashua Keynote address at Bosch ConnectedWorld Conference 2017

Share it with your friends Like

Thanks! Share it with your friends!

Close

Self-driving technology consists of processing of sensing data, the handling of HD-maps and the decision making processes required for merging safely into dense traffic – a process we call “driving policy”.

While processing sensing data is all about understanding and modeling the Present, Driving Policy is about modeling the Future.

Modeling the future raises new challenges of how machine learning should be incorporated to allow on one hand advanced human-like negotiation skills required for merging into dense traffic while guaranteeing functional safety.

Talk given at Bosch ConnectedWorld Conference 2017.

Prof. Shashua at World Knowledge Forum: Platform for Safe & Scalable AVs – https://youtu.be/7dLQQcR9kGM
Take a ride in Mobileye’s autonomous car: https://youtu.be/L1Bmc61l99A
Mobileye Three Major Pillars of Technology for Autonomous Car CES 2017 – https://youtu.be/yC7Bef_Mm-s
Autonomous Car Driving with Prof. Amnon Shashua – https://youtu.be/dhEgD6ZFlQE
Mobileye’s Autonomous Car – What the System Sees – https://youtu.be/jKfwHsHUdVc
Autonomous Car Technology – http://www.mobileye.com/our-technology/

Subscribe now to Mobileye on YouTube: https://bit.ly/2vRes7k

About Mobileye:
Mobileye’s advanced driver assistance systems (ADAS) technology is deployed in more than 50 million vehicles today and is integrated into hundreds of new car models from the world’s major automakers including Audi, BMW, FCA, Ford, General Motors, Honda, Hyundai, Kia, Nissan, Volkswagen, and more. Mobileye began with the vision of reducing vehicle collisions and resulting injuries and fatalities. Today, Mobileye makes one of the most advanced collision avoidance systems on the market, while working toward autonomous driving and the coming autonomous mobility-as-a-service (MaaS) revolution in road safety.


Connect with Mobileye:
Visit the Mobileye WEBSITE: https://www.mobileye.com/us/fleets/
Like Mobileye on FACEBOOK: https://www.facebook.com/Mobileye/
Follow Mobileye on TWITTER: https://twitter.com/Mobileye
Join Mobileye on LinkedIn: http://www.linkedin.com/company/mobileye

Mobileye® 8 Connect™ – Driven by Safety: https://www.youtube.com/watch?v=VbX0hJqOqSo

Comments

Jef says:

Does anyone know if the map light approach is also possible with lidar? Prof. Shashua gives the impression that this is not happening but I don't see way you couldn't limit the lidar approach to landmarks in the same way that they are doing with camera

Yuliang Li says:

Anyone know the title of the driving policy paper? Tried google but got no luck. Thanks.

imispgh says:

Excellent pitch

Simulation is key. And designs should not avoid the toughest things to do. Work on those up front.

RAND study showing that driving around gathering data as a primary means to designs and test these systems is extremely inefficient. 11 billion miles to show AP is ONLY 20% better than human drovers. Getting to 100% better is hundreds of billions of miles.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiAqf2zw4jTAhVH3IMKHexQB7oQFggcMAA&url=http%3A%2F%2Fwww.rand.org%2Fcontent%2Fdam%2Frand%2Fpubs%2Fresearch_reports%2FRR1400%2FRR1478%2FRAND_RR1478.pdf&usg=AFQjCNFrH_fndq2qvdtwQbtjHZvASk3Kvg

There are 3D radars. Might not be cost effective and small enough for a vehicle but they exist. See SPY-A used on Aegis ships for an example.

There are at least 2 rules in the double lane merge – thought clearly not very reliable. That being use of signals and car behind is supposed to yield the way. No different than single lane merge. Though clearly harder because there could be cars on each side parallel to each other etc.

Michael Micky מיכאל מיקי Ohana אוחנה says:

אלוף האלופים פרופ' אמנון שעשוע !

Jacques Shim says:

Great Amnon!

Tsvetomir Marinov says:

Productivity increases by using a combination of data collection and real-time processing by machine learning.
This means that to be effective the car must retain what is learned.
The statistical data from training (both "new" and "old") should be kept on each car.
How much "old" data you store on vehicle now ? 10GB, 100GB?
After 1000 km how much is the new data that is created?
How much data on board would be enough to make it acceptable for full autonomous driving?

Write a comment

*

Area 51
Ringing

Answer