Edge AI for Smart Cities: Enabling Real-Time Intelligence and Decision-Making

Abstract

Edge Artificial Intelligence (Edge AI) has emerged as a critical technology to make real-time decisions and rely on data driven intelligence in smart cities. The constraints of cloud-based frameworks are evidenced by the fact that urban environments generate huge volumes of sensor data, which necessitates decentralized intelligence. This paper gives a detailed discussion of the Edge AI architectures, the technologies behind them and how they can be used in smart urban systems. It analyzes such fundamental concepts as the edge-fog-cloud continuum, lightweight AI models, and distributed processing. Technologies that are facilitating these systems, including IoT sensor networks, 5G/6G connectivity, specialized AI hardware, and federated learning are addressed in terms of their effect on scalability, performance, and security. The paper discusses some of the important applications of smart cities such as smart transportation, citizen safety, smart utility, smart buildings, and environmental surveillance. Significant challenges, including the heterogeneity of devices, energy constraints, privacy concerns, and governance are examined and new research areas such as TinyML, digital twins, and quantum edge computing are examined. In general, this survey shows how Edge AI can transform the current state of smart cities to make them more responsive, resilient, and sustainable, allowing them to make autonomous decisions in real-time.