부산대학교 도서관

The proliferation of digitization in the automotive industry has witnessed the drastic transition from fossil-fuel vehicles to autonomous Electric Vehicles (EVs) to enable colossal data communication in an Industrial Internet of Things (IIoT) environment. However, the malicious attacker can deter message exchange between autonomous EVs with various security attacks such as ransomware, cyber-trojan, and injection. Thus, to mitigate the aforementioned security issues, we proposed a deep learning-based secure message exchange framework, i.e., SEAM to safeguard the data communication between autonomous EVs in an IIoT environment. The proposed framework leverages the Long Short-Term Memory (LSTM) model to classify message requests as malicious or benign, enabling seamless and secure communication between EVs. The message classification performed by the LSTM helps autonomous EVs decide the message's authenticity, reducing road fatalities. The model training includes various state-of-the-art deep learning models such as Gated Recurrent Unit (GRU), Recurrent Neural Network (RNN) and LSTM. Finally, the performance of the SEAM is evaluated considering various metrics such as F-score, precision, recall, loss curve, accuracy curve, and confusion matrix for the best performing LSTM model in which Adam yields the best performance over other optimizers.