Zonal Architectures to Make Fully Autonomous Vehicles Possible

The latest zonal architectures make it possible to have fully autonomous vehicles, using integrated system-on-chips (SoCs) connected through high-bandwidth in-vehicle networks. Zonal controllers can handle tasks such as processing camera, radar, lidar, and audio data, and then send this data to the central compute unit for the final sensor fusion. This requires a high-performance SoC that can process all sensor data while consuming the lowest possible power.

When it comes to autonomous vehicles, having a modular processor and AI platform that can effectively handle a wide range of workloads is crucial. It's essential to be able to scale computing power from 0.25 to hundreds of tera-operations per second (TOPS) efficiently. Advancements in automotive and AI technologies are made possible by the Tensilica DSPs, hardware accelerators, Neo NPU, and the latest SoC chiplet technologies.

cubes image

Next-generation zonal architectures enable fully autonomous vehicles

Comprehensive Solution

The only provider that offers compute IP, and integrated flow for chiplets, chip, board and system design.

ASIL-Ready

ASIL-ready compute IP, interface IP, and subsystems

Chiplet-Based SoC Design

Chiplet-based reference design and SW Development Platform for ADAS