Edge Computing and AI Integration for Low-Latency Decision-Making in Smart Cities and Industrial IoT

Abstract

The growing pace of smart cities and Industrial Internet of Things (IIoT) systems has amplified the need of smart low-latency decision-making systems with the capacity to handle extensive heterogeneous data volumes in real time. Traditional cloud-based systems typically experience communication latency problems, bandwidth constraints, and scaling issues, which adversely impact time-constrained applications like smart traffic management, industrial control, environmental sensors, as well as predictive maintenance. To overcome such drawbacks, this paper offers a combined Edge Computing and Artificial Intelligence (AI)-driven architecture to make low-latency decisions in small cities and industrial Internet of Things (IoT). The design presented is a hybrid of distributed edge nodes, localized AI inference engines, and cloud-assisted coordination aimed at assisting quick data processing on the proximity of the source device. The model of real-time anomaly detection and predictive analytics is a hybrid deep learning model consisting of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. Simulated smart city and industrial IoT data was experimentally evaluated with different network conditions. The proscribed framework was found to have 97.1% decision accuracy, latency of 61.8, and 34.5% better energy efficiency than the traditional cloud systems. The strongness and steadiness of the proposed framework was statistically approved on the 10-fold cross-validation. The findings reveal that edge-AI integration is an effective and scalable solution to next-generation intelligent infrastructures.