Intel Has a Go at Self-Driving CarsFebruary 14, 2017
Author: Mike Demler
Intel has accelerated its push into the self-driving car race by rolling out its new Go platform. Go combines in-vehicle computing and cellular connectivity with cloud data-center functions. The automotive hardware includes two development systems based on Atom or Xeon processors. These systems will connect to a private-cloud hardware/software platform for development, training, and updating of driverless vehicles.
The smaller of the two Go systems is an Atom-based version. Intel designed a new automotive-grade chip for this system, but it withheld details. It plans to certify the processor for ISO 26262 ASIL C. The Atom-based system can handle Level 1/2 ADAS functions such as adaptive cruise control (ACC) and collision avoidance, but the company is primarily targeting what it calls Level 3–plus systems with limited hands-free autonomous-driving capability.
For fully autonomous Level 4/5 driving, the high-end Go platform uses a pair of Xeon-powered CPU boards. The company withheld details of the next-generation Xeon automotive processors, but we expect they will resemble the configuration in its autonomous-vehicle system from a 2016 IDF presentation. That system combined an E5-series Broadwell processor with Arria 10 FPGAs.
Intel plans to attract automotive customers by offering an integrated software stack for training and updating vehicles; this stack will combine in-vehicle software with online and offline cloud data-center resources. By offering customers a consistent single-vendor tool set from the vehicle to the cloud, the company hopes to differentiate from competitors that lack such a broad software portfolio.
Subscribers can view the full article in the Microprocessor Report.
Subscribe to the Microprocessor Report and always get the full story!