Title: A Conceptual Model for AI-Driven Decision Systems in Autonomous Vehicles: Enhancing Edge Analytics and System Resilience
Authors: Wilfred Oseremen Owobu, Olumese Anthony Abieba, Peter Gbenle, James Paul Onoja, Andrew Ifesinachi Daraojimba, Adebusayo Hassanat Adepoju, Ubamadu Bright Chibunna
Volume: 9
Issue: 4
Pages: 236-262
Publication Date: 2025/04/28
Abstract:
The advancement of autonomous vehicles (AVs) has spurred a growing need for intelligent and resilient decision-making systems capable of operating under dynamic and uncertain environments. This paper proposes a Conceptual Model for AI-Driven Decision Systems that leverages edge analytics to enhance real-time processing, situational awareness, and overall system resilience in autonomous driving technologies. The model integrates artificial intelligence (AI), edge computing, and fault-tolerant design to improve decision accuracy and response time, addressing key challenges in latency, data overload, and system robustness. The proposed model features a layered architecture comprising perception, cognition, and execution layers, each embedded with AI algorithms optimized for distributed edge processing. At the perception layer, data from multiple sensors-including LiDAR, radar, and cameras-are pre-processed using lightweight convolutional neural networks (CNNs) at the edge. The cognition layer incorporates reinforcement learning and fuzzy logic systems for adaptive decision-making in complex and unpredictable scenarios, such as obstacle avoidance and dynamic path planning. The execution layer converts high-level decisions into control signals for actuators, ensuring smooth and safe vehicle operations. To address system resilience, the model includes redundancy protocols, real-time anomaly detection using recurrent neural networks (RNNs), and predictive maintenance strategies based on historical performance data. Additionally, the framework supports federated learning, allowing distributed AVs to share insights without compromising data privacy or bandwidth efficiency. Empirical simulations using synthetic driving datasets demonstrate enhanced system responsiveness, with a 32% improvement in decision latency and a 24% increase in fault recovery efficiency compared to traditional centralized architectures. These findings suggest that integrating AI with edge analytics and resilience mechanisms can significantly elevate AV decision systems' safety, efficiency, and scalability. This conceptual model offers a foundation for future developments in AV systems and sets a benchmark for intelligent, edge-enabled transportation technologies. It also presents a scalable approach for U.S. automotive and AI industries aiming to lead global innovation in autonomous mobility.