ADASConnected DriverExperts Say!Interviews

Artificial intelligence in ADAS: Mobileye

Published: August 16, 2016 

Shield+ is a wonderful example of the ongoing interaction between Mobileye, the provider of ADAS technology, and those enjoying the technology in the field.

How does vision based driver assistance compare with the Lidar/Radar based system for ADAS?

Mobileye essentially offers artificial intelligence. The optimal “machine” known for driving an automobile is still an alert and proficient human driver. Artificial intelligence (in the form of Mobileye vision technology) emulates the human driver and as such relies on our primary sense which is vision – that is the ultimate means for knowing “what is out there”, and of course a machine is not subject to distraction or other factors that are known to be the cause of accidents. We are a “Third Eye” in the car. Vision technology can see lane markings and street signs, vision technology knows what a pedestrian or cyclist looks like – and of course vision technology knows what is in our field of vision but nevertheless does not pose a risk of collision. This is the reason that it is the preferred solution. There are hybrid solutions that have been developed in order to augment vision, for example in inclement weather or poor lighting conditions, but the core solution will remain vision technology and this is the direction the industry has chosen.

Michael Hirsh_Mobileye
Michael Hirsh, Regional Manager for India Mobileye Aftermarket Division

Can you tell us about EyeQ SoC processing platform?

Mobileye has developed a family of chips called Mobileye EyeQ® to fulfill the need for low power and inexpensive computing platforms which are able to support computationally intensive vision applications and which meet automotive requirements. Initially the EyeQ powered the essential ADAS features such as Lane Departure Warning and Forward Collision Warning. As requirements from the OEM sector have developed, up to and including fully and semi-autonomous driving, Mobileye has continued to evolve the EyeQ platform in order to allow for simultaneous and instantaneous processing of the multiple inputs required to provide such capabilities.

What are your aftermarket solutions in ADAS?

Although initially dubbed “Aftermarket”, a more correct term – particularly in the Indian scenario – would be “Retrofit”. In a nutshell, Mobileye has taken the essential OEM-based ADAS features and bundled them in a platform that may be retrofitted in almost any vehicle on Indian roads, or rolling off the assembly line. These include Lane Departure Warning, Forward Collision Warning, Headway Monitoring & Warning (tailgating), Daytime Pedestrian and Cyclist Collision Warning and Speed Limit Indication. This retrofit platform is called Mobileye 630 and its implications for India cannot be overstated: World-class ADAS is available now to virtually the entire India road fleet – effectively “shortcutting” a decade-long evolution that has taken place in Europe, the US and other regions.

What is Mobileye Shield+?

Shield+ is a wonderful example of the ongoing interaction between Mobileye, the provider of ADAS technology, and those enjoying the technology in the field. What happened is that fleet operators and city planners came forward and said, okay – you are looking out the front of the vehicle, but what about the vulnerable road users that invariably located at the sides of buses and trucks, particularly in urban environments. They are effectively in the drivers’ blind spots and the results are all too known – a huge and tragic toll of pedestrians, cyclists and motorcyclists that are killed or injured during the turning and maneuvering of larger vehicles. In order to combat this, Mobileye provided additional sensors down the sides of large vehicles which deliver life-saving warnings to the driver that a vulnerable road-user is in  the path or the vicinity. In addition, these real-time warnings and their map location are recorded for later analysis by city planners, and these provide invaluable information on “hotspots” or those areas that effectively constitute “accidents about to happen”. Cities have always have had a map of where the crashes have occurred – now we provide a comprehensive map of where the accidents are likely to happen due to infrastructure deficiencies or other readily solvable factors.

What are the market indicators which show that a region is ready for adoption of ADAS?

ADAS is needed wherever there are road accidents. Today, in India, commitment is coming from the top. Road Transport and Highways Minister Nitin Gadkari has just this February appointed an eight-member taskforce to achieve a 50% reduction in road accidents in the coming five years. We have submitted our proposal in the framework of that endeavor and intend to play a pivotal role. India’s road toll is among the world’s highest and there is widespread and growing awareness that this this is a matter of national urgency, and indeed practical initiatives being taken on the ground. These are all clear indications that the time for large-scale adoption of ADAS in India is now.

How far are we in India from large scale adoption of ADAS?

We are already seeing the first practical steps being taken. The infrastructure for ADAS retrofit installation is already in India – first clients are large bus and truck fleets. And of course we are in active dialogue with government and regulatory partners.

Can we look forward to vision only based autonomous vehicles?

As I mentioned, vision technology will always serve at the core of autonomous solutions. We may expect radar and Lidar to be used to supplement sensing capability.

Whom would you consider competitor – Tier-1 companies working in ADAS or companies like Google, Uber and Tesla who are joining the automotive party?

Mobileye is regarded as the world leader and innovator in ADAS and works with leading Tier-1 companies, and through them with most leading auto manufacturers. There are additional players who occupy a relatively small part of the market whom we could consider competitors. We work with Tesla, and Uber and similar companies are potential clients. Yes, Google has demonstrated technology for autonomous driving. However, Mobileye has developed a mapping technology called REM™ (Road Experience Management™) which provides essential road data for autonomous vehicles with far less band width and far greater accuracy that other solutions on offer.

And a final word

I am thrilled and honoured to have the opportunity to participate in India’s ADAS drive, and even more so to have found here such willing and visionary partners for this worthy endeavour. These are exciting times and, given the pace of technological acceleration in India in all sectors, I believe Indians can look forward to enjoying lifesaving ADAS technologies in the immediate term. We are proud to be taking part in the charting of India’s road safety roadmap.

Michael Hirsh heads Mobileye’s aftermarket operations in India and is based at the company’s research and development headquarters in Jerusalem. His brief entails liaising with government agencies, and establishing the support infrastructure necessary for the propagation of Mobileye’s collision-prevention technology in all market verticals in India. Michael holds an MBA in Technology Management from Tel Aviv University.

Show More

Related Articles