At the recent Tech.AD conference in Detroit, Mich., automotive industry experts met to discuss various perspectives on how and when autonomous driving will be implemented, along with the potential concerns surrounding them. Concerns ranged from understanding the ethical dilemmas associated with automated driving to how to leverage insight from driver behaviors. There were many provocative topics discussed.
Micron’s addition to the discussion was the presentation Axel Schiller gave which hit upon the necessary amount of redundancy needed to ensure the safe functionality of autonomous vehicles and the cost tradeoffs associated with it. Axel drew comparisons to the airline industry, where the amount of redundancies needed depends on the system and what it is required to do. For example, 777s have multiple redundant navigation systems. In order to achieve level 5 autonomous driving how much redundancy do cars need? Axel estimated that in 2025 functional blocks needed in autonomous systems would run around $500.
If you add dual modular redundancy to the system, this amount doubles and that does not even contemplate the cost of memory added to the price. Experts are still working to figure out the amount of memory needed to support the sensors, gateway block, cognitive algorithm block and system host controller block. How much data will be stored in the car, how much will be offloaded to the cloud, and how much will be needed to handle the incoming V2X data.
We are continuing to work with other industry experts to find solutions for these issues, and we invite you to join the discussion! To learn how you can get involved, sign up for our automotive solutions updates or visit our Automotive Memory Solutions page.