Today, cars need to do more than stop and go quickly and comfortably. With data-storage requirements no longer just driven by navigation and infotainment systems, but by advanced driver-assistance systems, or ADAS (market research firm Strategy Analytics expects that by 2021, automotive OEMs will be spending more than $37 billion per year on assistance and safety solutions), data-storage subsystems must be developed with small packages to meet space constraints and high density to handle advanced applications.
It’s also no secret that automotive makers are racing against each other to produce the first autonomous cars, relying on radar, LiDAR, cameras, night-vision devices and other sensing technologies. As we move from level 1 and 2 autonomy toward level-5 full autonomy, it’s estimated there will be between 3 and 10 GB of new data generated every second. Even a car with level-2 autonomous capabilities generates up to 1 GB of data/second.
Since you can’t transmit everything to the network or the cloud (and you don’t want to for latency reasons), massive storage is needed (Fig. 1). This will result in a possible fourfold increase in non-volatile memory (NVM) content, increasing from 16 GB to 128 GB to 64 GB to 1.0 TB, depending on the level of autonomy.