Is it really hard to find rivals for Mobileye in the fields of ADAS and autonomous driving?

“We try to separate the concepts of the two revolutions, autonomous driving and assistance, and the revolution of saving lives can be achieved through the development of ADAS technology.”

Mobileye has always believed that in the process of developing autonomous driving, the development stage of assisted driving cannot be skipped, and assisted driving has huge potential.

As ADAS technology and autonomous driving become more and more mature, their role in ensuring driving safety is becoming increasingly prominent. However, there are still great challenges in evolving from ADAS to a true autonomous driving platform.

As an upstream enterprise in the automotive industry chain, semiconductor company Mobileye may not be well known to consumers. Its other identity is a subsidiary of Intel and also Intel's main force in the field of autonomous driving.

This article will take a closer look at Mobileye and learn about its progress and layout in ADAS, autonomous driving, etc.

Mobileye Overview

Mobileye was founded in 1999 by Professor Amoon Shashua and Ziv Aviram of the Hebrew University of Jerusalem with the goal of developing and promoting a vision system to assist drivers in ensuring passenger safety and reducing traffic accidents while driving.

Although Mobileye was founded in 1999, it was not until 2007 that the first car models equipped with Mobileye products were launched. It took eight years from research and development to commercial use. By the second half of 2013, Mobileye had sold 1 million chips. Mobileye successfully went public in July 2014. After the listing, Mobileye has been running wild, competing for the tenders of smart car safety equipment of major automobile manufacturers. By 2015, it had sold 10 million chips and had cooperation with more than 25 automobile manufacturers.

Almost all well-known automobile manufacturers are Mobileye's customers. Information on the company's website shows that a total of 17 automobile brands, including Audi, BMW, Citroen, Ford, GM, Honda, Hyundai, Jaguar, Land Rover, Nissan, Opel, Renault, Toyota and Volvo, have cooperative relationships with Mobileye.

In March 2017, Intel acquired Mobileye for $15.3 billion and merged the original autonomous driving division IDG with Mobileye. In less than two years after joining Intel, Mobileye quickly completed the integration with Intel's original autonomous driving department and formed a strategic layout of "taking root in Israel and radiating to the world", which accelerated the commercialization of Mobileye's autonomous driving.

Mobileye's business is mainly divided into two parts: front-end and after-sales. This year, the entire after-sales department and part of the front-end business team were integrated to form a new smart travel solution division IMS. The main reason for setting up a new department is that the development direction of the entire autonomous driving industry is no longer limited to OEM, but also has policy and regulatory considerations. Moreover, this technology will eventually be implemented, which also involves the issue of multi-scenario applications. Smart travel is an application that can cover the most touchpoints.

Machine vision technology has developed rapidly in recent years. Most companies attempt to enhance the ability of equipment to detect objects by improving the vision of cameras or adding radars. However, Mobileye has taken a different approach, using a single camera with attached sensors and unique algorithms to perform object detection tasks on a single hardware platform. This greatly simplifies the equipment installation process and greatly reduces costs, making it popular with major automakers.

The excellent market performance is due to Mobileye's clear path and rich and mature product family for autonomous driving and assisted driving.

ADAS system (EyeQ® chip)
is a camera-based advanced driver assistance system (ADAS). The range of ADAS systems lies in passive and active. Passive systems warn the driver of potential dangerous situations so that the driver can take corrective measures. In contrast, active safety systems take action automatically.

Passive systems include: Forward Collision Warning (FCW), which indicates that a collision is imminent given the current dynamics relative to the vehicle ahead. The driver is then required to brake to avoid a collision; Lane Departure Warning (LDW), which warns the driver of an unexpected/unindicated lane departure;

Active systems include: Autonomous Emergency Braking (AEB), which identifies an impending collision and brakes without any driver intervention; Lane Keeping Assist (LKA), Lane Centering (LC), Traffic Jam Assist (TJA), Traffic Sign Recognition (TSR) and Intelligent High Beam Control (IHC), among others.

If the vehicle path is slower, ACC automatically adjusts the host vehicle speed to a preset value (like standard cruise control). LKA and LC automate the vehicle to stay within lane boundaries. In traffic jam conditions, TJA is a combination of ACC and LC. It is these automated functions that form the building blocks of semi/full autonomous driving.

As a visual ADAS provider, algorithms and hardware are the core of Mobileye and also its advantages.

Mobileye's advanced driver assistance system mainly relies on three parts: camera components (camera, speaker, main chip), EyeWatch (display) and OBD junction box. Among them, Mobileye's camera is the core part of the whole system design, using the specially customized EyeQ visual processing chip. It can be said that under absolute environment, one camera can realize the above-mentioned very comprehensive assisted driving functions.

Mobileye's technology has had a huge impact on the entire automotive industry. Since 2007, Mobileye's EyeQ® technology has been installed in cars and has begun to succeed in the market. It has been adopted by 27 automakers. What distinguishes the EyeQ® series of system-on-chip (SoC) devices from competitors is that it can support complex, high-computation visual processing while maintaining low power consumption even when installed on the windshield.

In 2015, Mobileye released the fourth-generation EyeQ4 visual processing chip. The EyeQ4 chip uses 14 computing units, 10 of which are special vector accelerators. The benefit is that it greatly improves the performance of visual processing and data interpretation, and can achieve 2.5 billion operations per second. The new product Mobileye 8 Connect™ equipped with the EyeQ4® chip has been gradually applied to new vehicles since 2018. Mobileye 8 Connect™ is Mobileye's new generation of aftermarket collision avoidance system, which can better collect road traffic and surrounding environment data, and then generate high-precision maps. It collects dynamic data based on EyeQ4®, making cities smarter, safer and ready for autonomous driving. It can be seen that EyeQ4 has once again upgraded Mobileye's autonomous driving assistance system, and with the support of Intel, it has made the commercialization of L4-level autonomous driving possible.

Mobileye is currently developing its fifth-generation SoC, EyeQ5, which serves as the central visual computer and performs sensor fusion for fully autonomous (Level 5) vehicles that will hit the road in 2020. To meet power and performance targets, EyeQ® SoCs are designed on the most advanced VLSI process technology nodes, with the fifth-generation EyeQ5 having a 7nm FinFET and delivering ten times the performance of EyeQ4®.

In the future, Mobileye will open up many technologies, specifically in hardware and software. At this year's CES, Mobileye announced the opening of the EyeQ5 platform, which is a major breakthrough that can attract more partners. In fact, Mobileye has always provided core chips to Tier 1. Most of the re-integration after ADAS calculations or autonomous driving sensor calculations are completed together with Tier 1 and car manufacturers. In this process, Tier 1 or car manufacturers will do secondary development based on Mobileye's core algorithms.

In 2018, shipments of EyeQ® chips increased by approximately 42% year-on-year compared to 2017, and over the past decade, more than 30 million cars powered by Mobileye's leading technology have been on the road.

Road Network Collection Management™ (REM™)
In Mobileye's view, in the process of L2+~L4 autonomous driving evolution, whether the map can be updated in real time will play a key role. In the commercialization process, cost-effective and affordable sensors and map update data that can be updated quickly and cost-effectively are more needed. The Intelligent Mobility Solutions Division (IMS) established by Intel this year has cooperated with major cities around the world to carry out the layout of map updates.

Mobileye's advantages lie in its know-how accumulated over the years in ADAS R&D and its key REM map technology (Road Experience Management). In addition to autonomous driving, REM can improve the existing ADAS user experience at minimal cost and provide real-time data for smart cities, facilitating automatic infrastructure surveys for urban planning.

Currently, Mobileye has completed a number of business arrangements related to map updates around the world, including various pilot projects to prove the effectiveness of the technology, such as making maps with real-time update capabilities for various cities in priority to attract car companies and autonomous driving startups to come for road tests.

At this year's CES, the concept of L2+ was frequently mentioned by many manufacturers, and Mobileye also announced that its cooperation with Volkswagen* has reached the L2+ level. For Mobileye, while self-driving cars are on the road, Mobileye will also devote some of its energy to assisted driving, but if it wants to achieve the vision of saving lives, the current assisted driving is far from enough, so L2+ with more possibilities was born.

 
Mobileye's L2+ has further improved the L1 to L2 assisted driving functions and provides more convenience for today's drivers. L2+ improves the vehicle's semi-automatic driving capabilities by leveraging Mobileye's Road Network Collection Management™ (REM™) map solution and Roadbook™ data to achieve adaptive cruise control, lane departure correction and other higher-level control functions in more complex driving scenarios such as urban roads. The improved functions not only narrow the gap with higher levels of automated driving, but also significantly improve driving convenience and safety.