Home > AI Solutions > Artificial Intelligence > White Papers > Dell Automotive Reference Architecture > Overview
Automotive OEMs, automotive tier 1 suppliers, startups, rideshare, military, trucking, mining, and other Advanced Driver Assistance Systems/Autonomous Driving (ADAS/AD) programs are advancing from SAE levels L2 to L5. Customers are becoming increasingly overwhelmed by the amount of data that must be ingested, cleaned, trained, verified, simulated, updated, and stored. They are also feeling the competitive pressures to reduce development cycle time without sacrificing performance or increasing costs.
The Dell Automotive Reference Architecture (DARA) offers a solution to this challenge. The fundamental piece of DARA is an AI “starter” or “base” kit. Because of its modular nature, “scalable units” (SU) of infrastructure that consists of compute, networking, storage, or simulation platforms can be added later as the environment scales. DARA also defines a software infrastructure stack for AI model creation and development. This software infrastructure stack is extended to support AI model validation and verification (V/V). AI model V/V includes software-in-the-loop (SiL) and hardware-in-the-loop (HiL) simulation environments for L2 to L5 autonomous vehicles. These environments are architected to address the problems of scale without sacrificing performance or business continuity. They are targeted for on-premises or hybrid cloud environments or consumed as-a-Service (aaS) for automotive OEMs/Tier 1s, robotaxis, automotive startups, mining, and military applications.