EXPLAINABLE AI FOR CRITICAL DECISION MAKING IN AUTONOMOUS VEHICLES USING FPGA-AND GPU-BASED HIGH-PERFORMANCE COMPUTING SYSTEM DESIGN

This paper explores the integration of Explainable Artificial Intelligence (XAI) with FPGA- and GPU- based High-Performance Computing (HPC) systems for critical decision-making in autonomous vehicles. Autonomous vehicles hold great promise in revolutionizing transportation, but ensuring their safety and reliability in complex real-world scenarios remains a challenge. XAI offers an avenue to achieve transparent and interpretable decision-making, instilling trust and confidence in AI-driven systems. Leveraging the parallel processing capabilities of FPGA and GPU architectures, our proposed approach enables real-time, highspeed, and energy-efficient data analysis. The research presents a framework that seamlessly merges XAI principles with FPGA- and GPU-based HPC, optimizing the hardware-software interface for efficient AI model deployment. Simulations and real-world experiments demonstrate the superiority of our approach in enhancing autonomous vehicles’ decision-making while maintaining interpretability. This work contributes to building a transparent and accountable AI-driven mobility ecosystem, fostering wider acceptance of self-driving vehicles.