•  
  •  
 

Abstract

Port and vessel networks increasingly operate on IP/Ethernet backbones with high‑noise, high‑dimensional traffic. We present a lightweight hybrid intrusion‑detection model that couples a variational autoencoder (VAE) with a multilayer perceptron (MLP) and augments training with a boundary‑oriented latent‑space mixup strategy. The VAE models the distribution of normal traffic and identifies anomalies through reconstruction errors. Subsequently, it generates robust latent vectors, enabling the MLP to perform highly accurate supervised classification. On the UNSW‑NB15 dataset, the proposed pipeline attains ≥97% accuracy and an outstanding recall of 99.56% in binary intrusion detection, and visualization of the latent space (PCA) together with reconstruction‑error analyses and the confusion matrix corroborates clear anomaly separation. The model uses a compact design. The VAE produces latent representations, and an MLP classifies them. With latent-space mixup, it detects new anomalies and reliably classifies known threats, making it suitable for smart-port monitoring and ship-communication security at the edge.

Abstract Image

Download
COinS