SDV-Aware Evolved Automotive Safety Architecture

SDV-Aware Evolved Automotive Safety Architecture

연구 분야: Verification

학회: 2024 IEEE 5th Women in Technology Conference (WINTECHCON)

With increasing automotive industry demands for electronic controls around infotainment, powertrain, chassis and body control, there is significant rise in electronics in vehicles. Additionally, there are requirements for Over-The-Air (OTA) software (SW) updates in existing hardware (HW). Hence automotive industry is evolving towards a Software Defined Vehicle (SDV) concept. With the rising expectations of modular and independent zonal processing, traditional functional safety architecture doesn't suffice. In this paper, we describe an evolved architecture of functional safety to keep up with the general evolution of automotive requirements. It utilizes a modular and hierarchical safety architecture with context-aware safety management and enhanced diagnostics. In this distributed, de-centralized safety architecture, the faults in one domain (or zone) of the design do not end up propagating to all domains (or zones) of vehicle. Deployment of this architecture on a state-of-the-art automotive SDV solution shows that 90% of the faults can be handled using software mechanisms compared to 25% on traditional safety architectures (with the remaining 75% ending up as a hardware reset). This leads to improvements of several-orders (typically 10X-100X) around recovery time and system availability.