Live Demonstration: AI-based System Latchup Detection and Protection for COTS Systems

The adoption of Commercial Off-The-Shelf (COTS) ICs in modern satellites faces challenges from radiation-induced latchup events, with current protection methods showing significant limitations in detection accuracy and applicability. This demonstration presents a novel adaptive AI-based latchup detection and protection system featuring two-stage training and LSTM neural network analysis. Our FPGA implementation achieves 90% detection accuracy without extensive pre-characterization. Visitors can validate the system's performance through real-time interaction with various latchup scenarios.