Presentation Information
[C-10_C-11-06]Advanced Engineered Substrate Technologies Enabling Functional
Safety and Cyber-Resilience in Next-Generation Mobility AI
〇Toshiaki Fujita1 (1. SOITEC Japan Inc.)
Keywords:
FD-SOI,Power-SOI,Functional Safety,Cyber-resilience,Laser Fault Injection,Edge AI,Autonomous Driving
As autonomous driving and centralized zonal architectures expand, the demand for robust Edge AI (Physical AI) in next-generation mobility systems scales exponentially. Implementing high-performance computing within stringent automotive power budgets while satisfying strict Functional Safety (FuSa) and cyber-security standards poses significant hardware challenges. Traditional bulk silicon technologies increasingly confront immutable physical limitations, including high-temperature thermal runaway, latch-up vulnerability, and susceptibility to physical fault attacks. This paper evaluates fully depleted silicon-on-insulator (FD-SOI) and Power-SOI engineered substrates as essential foundational platforms to overcome the bulk silicon reliability wall. By leveraging absolute dielectric isolation and dynamic body-biasing, these technologies inherently provide intrinsic latch-up immunity, mitigate high-temperature leakage, and have been shown to suppress LFI susceptibility by up to 150$\times$ under reported experimental conditions, providing a comprehensive hardware-level perimeter for secure, dependable mobility AI.