January 22 – By Sergio Marchese
Advanced electronic systems for connected autonomous vehicles (CAVs) and other safety- and security-critical applications use complex software stacks. At the bottom of the stack are integrated circuits (ICs) that include general-purpose and workload-optimised processing engines, and other semiconductor intellectual properties (IPs). However, hardware vulnerabilities may compromise the entire system.
Ensuring that ICs, both ASICs and FPGAs, have high integrity requires adequate hardware-development flows that deliver evidence of functional correctness (the IC satisfies its intended mission requirements), safety (the IC can prevent or control failures that could occur during operation due to physical effects), and trust and security (the IC doesn’t include unexpected or malicious additional functions that could be exploited in cyberattacks). Integrity properties aren’t an afterthought. All IC and IP development stages, including pre-silicon validation and verification, need adequate tools and methods to achieve high integrity. The safety and privacy of people is at stake.
Proving functional correctness of complex hardware designs is challenging. Over the past 20 years, new electronic-design-automation (EDA) technologies and methods have emerged to address this task. Safety requirements were once confined to niche, low-complexity applications. In the past 10 years, this has changed dramatically.
With the advent of advanced driver-assistance systems (ADAS) and developments toward self-driving cars, new and established IP and IC providers have deployed functional safety flows for the development of complex electronic systems. Much credit goes to the ISO 26262 functional safety standard for road vehicles, which has enjoyed widespread adoption since its first draft was published more than 10 years ago.