A Fog-Enabled OMNeT++ Framework for Network Anomaly Detection in Smart City Environments

In the age of smart cities, maintaining safe and reliable data transmission over large-scale sensor networks is becoming more important. These networks, which are often made up of smart cameras and IoT devices, are susceptible to abnormalities like rogue nodes, sensor failures, and traffic spikes, which may jeopardize system integrity and performance. Traditional centralized anomaly detection methods suffer from latency and scalability difficulties, particularly in high-density settings. To solve this issue, this paper proposes a new anomaly detection system that combines Ant Colony Optimization (ACO) and clustering approaches in an OMNeT++ simulation environment. The system uses a fog-cloud architecture, with fog nodes doing localized processing and clustering to decrease latency and data overhead, and ACO for efficient data routing and anomaly detection. Simulations were run with three distinct sensitivity settings: baseline, high, and low, to assess detection accuracy, precision, recall, and F1-score. The suggested technique showed considerable increases in all measures, with baseline scenario accuracy improving from 86.5% (without detection) to 92.4% (with detection), and F1-score from 0.64 to 0.837. Furthermore, the system demonstrated improved processing efficiency and decreased network use at both the fog and cloud levels, indicating its applicability for real-time anomaly identification in dynamic and scalable smart city networks.