EP 34: Autonomous cars: Why don’t we have self-driving cars yet?

Summary:

In this episode, we discuss the challenges and current state of autonomous vehicle development, focusing on the role of advanced semiconductor chips. We explain the different levels of autonomous driving automation, from Level 0 (no automation) to Level 5 (full automation). We highlight the need for increasingly sophisticated chips with smaller nanometer (nm) sizes to power these advanced systems, especially for achieving Level 5 autonomy. While progress is being made, we note that standardized road conditions remain a significant obstacle to widespread deployment of fully autonomous vehicles.

Questions to consider as you read/listen:

1. What are the technological hurdles preventing widespread adoption of self-driving cars?
2. How does the current state of semiconductor technology impact the development of autonomous vehicles?
3. What are the key differences between the various levels of autonomous driving automation?

Long format:

Autonomous cars: Why don’t we have self-driving cars yet?

The current conventional nomenclature for autonomous vehicles is from the SAE.

The six levels of autonomous car automation are:

Level 0: No driving automation

Level 1: Driver assistance

Level 2: Partial driving automation

Level 3: Conditional driving automation

Level 4: High driving automation

Level 5: Full driving automation

It all depends on what level you want. As a general rule, the higher the level the lower the nm on the chip. That’s simplistic because there is a lot that goes into it beyond the chip such as hardware (LIDAR, radar, sensors, etc). But simplistically that is so.

Most automotive semiconductors are based on nodes below 22 nm. However, some vehicle designs use leading-edge 7- and 10-nm semiconductor chips, and 5-nm designs are in development.

The next generation of chips for autonomous driving systems will use technologies closer to the cutting edge, in the 1-5 nm range.

China at best is at 7nm currently. They are developing but not yet deployed 5nm. As I have written all over this community with lots of words, I don’t see China cracking into the sub 5nm reality for some time.

The current state of the art is 3nm with IBM at 2nm. No one is at 1nm but bold prediction at the 2025 conference I bet it will be revealed.

The bane of the existence of the dream of automated cars is in non standardized road conditions and non uniform lane markers. There is a great story of how Elon Musk told his engineers that he demanded to drive in a fully automated Tesla from his house to his headquarters with zero human intervention. It was a demanding timeline. For almost all hours the engineers tried and tried and tried and failed….. until they decided to go outside the box, and repair and repaint all of the roadway between his house and the HQ. Problem solved. Elon’s demand accepted. It’s a good story but a better illustration of the issues with the goal of achieving level 5 cars that will function in the wild. China is a country that would repaint all of its roads and make it supremely uniform perhaps. Will the US? Dunno. I would not bank on it.

Sources:

https://digital.bnpmedia.com/publication/?i=667937&article_id=3726579&view=articleBrowser

https://www.embedded.com/buying-autonomous-electric-vehicles-will-soon-be-just-like-buying-a-laptop/#

https://www.synopsys.com/blogs/chip-design/autonomous-driving-levels.html#

https://www.synopsys.com/blogs/chip-design/autonomous-driving-levels.html#