Real-Time Anomaly Detection in Smart Grids,Industrial IOT,Autonomous Vehicles.

The rapid expansion of cyber-physical systems such as smart grids, Industrial Internet of Things (IIoT),and autonomous vehicles has increased the demand for reliable, real-time anomaly detection systems.These systems are critical for maintaining operational safety,reducing downtime, optimizing energy usage,and preventing cyberattacks. However,traditional anomaly detection approaches suffer from high latency,limited scalability, and lack of adaptability to dynamic environments.This paper proposes an integrated real-time anomaly detection framework using a hybrid approach combining Edge AI,recurrent neural networks (RNN-LSTM), and unsupervised clustering (Isolation Forest). The model is deployed on edge devices for low-latency processing and uses a multi-stream data fusion mechanism to analyze sensor,communication, and control signals simultaneously.Experimental simulations across smart grid load datasets, IIoT factory sensor datasets,and vehicular CAN-bus datasets show detection accuracies of 96.4%, 94.8%, and 97.1%,respectively.The proposed solution demonstrates superior performance in terms of detection speed, false-positive rate, and adaptability compared to traditional systems, making it suitable for high-reliability real-time environments.