See you after CES, all-star lineup of autonomous driving technology from China and the United States | Countdown to the summit

In the past two or three years, CES has become an important battlefield for automotive technology. In 2015, Mercedes-Benz showcased the intelligent concept car F105 Luxury in Motion, and in 2017, Toyota showcased the artificial intelligence concept car concept-i at CES. Between the two stunning concept cars, major technology companies, automobile manufacturers, and technology suppliers are investing heavily in technology development, and autonomous driving is developing rapidly. In the blink of an eye, the autonomous driving described by Mercedes-Benz F105 and concept-i is very close to reality in terms of technology.

After the upcoming CES 2018, New Intelligent Driving will hold the GAIR Silicon Valley Intelligent Driving Summit in Silicon Valley, the headquarters of American technology (tickets for the summit are currently open for booking, and Christmas tickets are expected to end at 23:59 Pacific Time on January 6, 2018 https://gair.leiphone.com/gair/gairsv2018 ).

Silicon Valley is not only a gathering place for the world's top technology companies and technology talents, but also a major center for the development of autonomous driving technology. Looking at the autonomous driving road test licenses issued by California alone, 47 American and multinational companies have obtained the qualification to test on public roads, which means that at least 100 autonomous driving vehicles are tested on public roads in Silicon Valley and various towns in California all year round.

The 47 companies include:

• Google (Waymo) and Apple, the two most valuable technology companies in the world;

• The three giants in the consumer electronics and chip industries: Nvidia, Qualcomm, and Samsung Electronics;

• The three largest automakers in the U.S.: General Motors (Cruise Automation), Ford, and Chrysler (in partnership with Waymo);

• Japanese automakers, including Honda, Nissan, and Subaru, have even surpassed local automakers in the U.S. market share;

• The three German automakers are Volkswagen, Mercedes-Benz and BMW;

• The rapidly developing Chinese automakers, including SAIC Motor, one of the Fortune 500, and Changan Automobile, which recently obtained a road test license;

• Emerging car companies, including Tesla, Faraday Future, and NIO;

• There are also world-class technology suppliers such as Bosch, Continental, Delphi, and Valeo;

• In addition to the above, there are about 20 startups of all sizes that represent the innovative power of Silicon Valley.

And this is just the tip of the iceberg of Silicon Valley's autonomous driving industry. According to incomplete statistics, there are more than 300 autonomous driving startups in Silicon Valley alone, and more than 1,000 companies related to the industry chain. For example, the autonomous driving technology service provider AutonomouStuff told us that they currently serve more than 2,000 customers.

Since 2016, Leiphone.com New Intelligent Driving has been widely reporting on autonomous driving technologies in China and around the world. Since March 2017, Leiphone.com New Intelligent Driving has been focusing on reporting on the technology and application progress in various fields of autonomous driving in Beijing, Shenzhen, Shanghai, Silicon Valley and other places. On January 16, 2018, Leiphone.com will invite the top autonomous driving technology teams from China and the United States to hold the GAIR Silicon Valley Intelligent Driving Summit in Silicon Valley. Here are the highlights of the summit:

1. Top universities in North America, the birthplace of autonomous driving

The DARPA Grand Challenge, which has been held since 2004, is the origin of the global industrialization of autonomous driving technology. From 2004 to 2007, the DARPA Grand Challenge was held three times, with a prize of up to one million yuan. From autonomous driving in desert roads to autonomous driving in urban roads, a large number of autonomous driving talents have been trained in universities in the United States that are good at computers and robotics, including the CMU Carnegie Mellon team, the Stanford team, the MIT team, the Cornell team, and so on.

To this day, members of these teams are still the most important developers and promoters in the field of autonomous driving. Former team members have been associated with and founded autonomous driving teams such as Waymo (Stanford Team), Uber ATG (Carnegie Mellon Team), Nutonomy (MIT Spinoff), etc.

*Stanford Stanley Chariot

At the 2018 GAIR Silicon Valley Intelligent Driving Summit, New Intelligent Driving also invited these earliest DARPA participants, including Adrian Kaehler, a member of the 2005 DARPA Stanford team. He designed the machine vision system for the tank Stanley at the time and assisted the team in winning the 2005 DARPA Challenge. It was also based on the DARPA Challenge and the development of Stanley that Willow Garage, one of the most important companies in the history of Silicon Valley robotics, was later derived. After the DARPA event, Adrian Kaehler later became a core member of OpenCV, the founder of Deep Learning Group, and the vice president of Magic Leap.

In addition to the Stanford team, Xinzhijia will also invite speakers from the early DARPA Grand Challenge CMU team and Cornell team to review with us the glorious years when the autonomous driving industry was in its infancy.

2. Cutting-edge sensor technology and the infrastructure for autonomous driving

Sensors are the first link in the autonomous driving technology chain of perception-recognition-decision-control. In 2018, Waymo, Uber, Lyft, Baidu and other companies have set their sights on the normal operation of autonomous driving projects, and the mass production of sensors has become a basic condition. Today, all autonomous driving developers on the market are still looking for a low-cost, high-resolution, stable, and mass-producible sensor. At this GAIR Silicon Valley Smart Driving Summit, you will see:

*Velodyne solid-state lidar Velarray prototype

Velodyne CTO Anand Gopalan, Velodyne is the undisputed leader in the global LiDAR field. Since the DARPA Grand Challenge, Velodyne's 64-line LiDAR has been almost the standard for autonomous driving fleets. By the end of November this year, they had officially announced a LiDAR with up to 128 lines. In addition to Velodyne's mechanical LiDAR, which occupies most of the global market share, Velodyne is also developing a new type of LiDAR that has no mechanical rotating mechanism, is solid-state, easy to produce, and meets automotive standards. Anand Gopalan is the core leader in the development of Velodyne's new LiDAR.

*Innovusion's test vehicle and the lidar prototype on the roof

In addition to Velodyne, there is still a lot of room for innovation in the field of LiDAR. Innovusion is a sensor and sensor fusion solution company headquartered in Silicon Valley. This month, they just released a sensor that integrates cameras and LiDAR at the hardware and software levels. After the demo stage, their mass production goal is that LiDAR can detect objects with a reflectivity of only 10% at a distance of 150 meters and form a dense point cloud effect (equivalent to 300 mechanical lines). Each pixel in the point cloud carries RGB information and 3D coordinate information. The overall solution cost is controlled at US$1,000. Innovusion CEO Bao Junwei will share their sensor technology at this summit.

*Metawave's mmWave radar demonstration

Metawave is a developer of a new type of millimeter-wave radar. Compared with other vehicle-mounted sensors, millimeter-wave radar is not affected by light, works 24 hours a day, and is inexpensive. However, for autonomous driving, the resolution of millimeter-wave radar is too low and the information obtained is too little. The technology developed by Metawave is used to solve this problem. In 2006, Maha Achour founded Rayspan, a metamaterial research and development company, as CTO. The company's main technology is to use metamaterials to develop a precision and controllable antenna.

Eleven years later, Maha Achour, together with his current company Metawave’s CTO Bernard Casse and Engineering VP George Daniel, once again applied metamaterial antennas to the field of automotive millimeter-wave radar. Metawave’s millimeter-wave radar technology features the ability to emit highly directional and controllable electromagnetic beams, thereby achieving directional detection and scanning of millimeter-wave radars, and accurately detecting the position and speed of targets.

3. Key links of autonomous driving are gathered together to build a complete technical system

Autonomous driving is a huge project across multiple fields. In addition to sensors, we also need:

*Two processors released by Horizon Robotics in Beijing

In-vehicle high-performance, low-power computing center

In 2015, Yu Kai resigned from his position as vice president of Baidu Deep Learning Institute and founded Horizon Robotics. One of Horizon Robotics' main businesses is to provide autonomous driving solutions that combine high-performance algorithms with low-power chips. In 2016, Horizon Robotics was already a supplier to Bosch, a top automotive Tier 1 supplier. On December 20, Horizon Robotics released two self-developed chips, the Journey and the Rising Sun, in Beijing, specifically for intelligent driving and camera applications.

*AutoX's self-driving prototype

Accurate and efficient CV algorithm

At present, CV has no way to solve all the problems of autonomous driving, but Xiao Jianxiong, the founder of AutoX, believes that "people underestimate the camera, which leads to a lack of effort to bring out its functions. There is still a lot of room for improvement, but people just don't work hard enough to make use of it." Six months after its establishment, AutoX quickly developed an autonomous driving prototype car with cheap cameras as the main sensor solution.

High-precision maps that machines can understand and describe the physical world

Self-driving cars need to understand the world around them and rely on information obtained from sensors to model the perceived world in real time. However, a better way is to have an established accurate model of the world in the car or in the cloud. Therefore, the higher the data accuracy required for high-precision maps, the better. The difficulty is that such maps are very difficult to store and maintain. DeepMap was founded based on this opportunity. The founding team of DeepMap are all members who have been immersed in the map industry for many years and have a deep understanding of the map operation architecture.

*Righthook.io’s simulation environment based on high-precision maps

A simulation system that can complete billion-kilometer and 10-billion-kilometer level tests online

Real-world autonomous driving tests cannot be repeated and controlled, so autonomous driving requires a testing framework that can quickly verify the reliability and safety of each iteration of the system without testing in the real world. What RightHook does is simulation based on high-precision maps. It does not rely on any previous sensor data, but creates all data in the simulation. They reconstruct the entire simulation environment based on the high-precision maps used by various autonomous driving companies to properly reproduce the real world.

At the GAIR Silicon Valley Smart Driving Summit, you can clearly understand all the key links mentioned above.

4. China-U.S. open autonomous driving platform to accelerate technology development

Solving the problem of autonomous driving now requires new sensors, computing units, algorithms, data, etc. But in the end, most of the problems of autonomous driving can be solved by accumulating data to promote algorithm evolution. Different driving environments around the world will form different technology alliances or data alliances. Mobileye is building an alliance through integration with Intel, which is a semi-closed technology platform. There are more open technology platforms in China and the United States, such as Lyft in the United States and Apollo in China.

In April this year, Baidu Apollo announced that it would be open to the public. In July, Baidu opened Apollo 1.0. In September, Apollo was updated to 1.5. At CES 2018, Baidu will open the capabilities of Apollo 2.0, which is said to be able to achieve autonomous driving under simple urban road conditions. This year, Apollo has successively formed alliances with top suppliers such as Bosch and Continental, most domestic automakers and Tier 1.

In addition to large companies, new technology companies also have a platform vision. Renovo.auto is committed to building an open autonomous driving operating system for on-demand travel. This month, Renovo.auto announced a partnership with Samsung Electronics, whose first autonomous driving fleet deployed in California will be equipped with their autonomous driving operating system.

At this summit, you will see open ideas and R&D experience sharing brought by different platforms.

*Renovo's prototype car for demonstrating autonomous driving operating system

5. The most aggressive automakers in mass production of autonomous driving

Of course, autonomous driving will eventually be implemented in cars. American automakers are actively developing autonomous driving. General Motors Cruise's autonomous driving taxi service is about to be launched, Ford has invested in Argo and invested heavily in research and development in San Francisco and Pittsburgh, and Chrysler has in-depth cooperation with Waymo.

In another of the world's largest automobile consumption markets, China's unique road and regulatory environment will bring completely different market opportunities to automakers.

In 2014, SAIC established a technology venture capital fund, SAIC Venture Capital, in Silicon Valley. In three years, they have invested in 22 projects in the United States alone. In 2015, SAIC established the SAIC USA Innovation Center to recruit talents in Silicon Valley to carry out cutting-edge research such as autonomous driving. SAIC is also the first domestic veteran automaker to obtain autonomous driving testing qualifications in California.

On the other hand, new car companies based on electrification and intelligence are also not to be outdone. Earlier this month, Xpeng Motors' 1.0 version of mass-produced cars hit the road in Beijing, and its autonomous driving R&D team in Silicon Valley is also developing its own autonomous driving operating system for two models that will be mass-produced in 2018 and 2019. Gu Junli, VP of autonomous driving at Xpeng Motors, said, "The essence of new cars lies in intelligent driving." In the next few years, car companies will continue to upgrade and iterate autonomous driving technology to compete in a differentiated manner, which will test their system design capabilities, software iteration capabilities, and even hardware R&D capabilities.

The above are some of the guests at the GAIR Silicon Valley Smart Driving Summit on January 16, 2018. Dozens of top autonomous driving teams from China and the United States will give speeches at the summit. In addition, nearly 100 companies in the autonomous driving industry chain will communicate on site.

For more information, please visit the following links Lock in Christmas Special Tickets:

gair.leiphone.com/gair/gairsv2018

Part of the confirmed guest lineup: