We can count at least 10 ADAS/autonomous vehicle platforms. How will these hardware and software platforms overlap, interact and compete?
Most non-gearheads who have grown up outside Detroit have some difficulty understanding the inner workings of the auto industry, particularly how car OEMs, tier ones and component vendors mesh together to get millions of cars off the assembly line.
This choreography is now complicated by technology suppliers — none based in Detroit and all new to the traditional automotive market. These players include Tesla, Waymo (formerly Google) and chip companies like Nvidia and Intel.
In an industry jumbled by newcomers, the partnerships among OEMs and tier ones are no longer business as usual. Existing alliances might be suddenly superseded by new ones. Competing technologies within rival autonomous vehicle platforms render the whole landscape alien and unpredictable.
The reporter’s life, of course, would be so much easier if the industry’s battle over automated vehicle platforms could be simply framed as a race between Nvidia vs. Intel. But nothing could be further from the truth.
On one hand, Nvidia wants the world to believe advancements in deep learning are swiftly and vastly transforming the automotive industry.
On the other hand, Texas Instruments, which has put more than 150 million of its ADAS and digital cockpit SoCs on the road, is confident that it can push its current [Level 1 & 2] platform to support Level 3 automated vehicles later this year. TI aims to take customers to the next level while protecting its software investment. TI also claims that a low-power embedded solution for running an AI-inference model is also in the offing.
The two companies represent polar opposite views and approaches, about how fast and how drastic the automotive industry is changing.
ADAS/Autonomous Vehicle (AV) platforms
“The auto industry has historically been a huge networked ecosystem with multiple tiers of suppliers,” Mike Demler, a senior analyst at the Linley Group, reminded us. “No one company can supply everything.”
Demler believes that at this stage, “most carmakers are still evaluating and developing. No one is looking to buy a complete stack.”
However, we can count now at least 10 ADAS/autonomous vehicle platforms — publicly touted and promoted by different parties.
Automated driving platforms we’ve seen so far include:
As shown in the table, not every AV platform can handle the gamut of ADAS and autonomous functions. Even Nvidia’s AV platform, focused on deep learning, can be described as a partial solution. It “will still need a safety monitor that can process control signals and monitor that,” noted Phil Magney, founder and principal advisor for Vision Systems Intelligence (VSI).
Egil Juliussen, director, research, infotainment & ADAS at IHS Automotive, told EE Times, “I think there will be multiple software and hardware platforms that overlap, interact and compete with each other.”
Next page: AV Stacks