CyberTrace IOT: Attack Detection and Attribution in IoT-Integrated CyberPhysical Systems

Abstract

The rapid proliferation of Internet of Things (IoT) devices integrated into Cyber-Physical Systems (CPS) introduces significant cybersecurity challenges. Traditional intrusion detection systems fail to address the heterogeneity, resource constraints, and real-time requirements of IoT-CPS environments. This paper presents CyberTrace IOT, a novel framework for attack detection and attribution in IoTintegrated CPS. CyberTrace IOT employs a hybrid deep learning architecture combining Long Short-Term Memory (LSTM) networks with Graph Neural Networks (GNN) to model both temporal behavioral patterns and inter-device dependency structures. The proposed system achieves a detection accuracy of 96.8%, precision of 95.7%, recall of 95.2%, and an F1-score of 95.4% on benchmark IoT attack datasets. Experimental results demonstrate superior performance over state-of-the-art baselines including Decision Tree, Random Forest, SVM, and standalone LSTM models. The attribution module identifies the attack origin with 93.6% accuracy, enabling rapid incident response in real-world deployments. Keywords: IoT Security, Cyber-Physical Systems, Intrusion Detection, Attack Attribution, LSTM, Graph Neural Networks, CyberTrace IOT