Deep Learning-Based Intrusion Detection for Automotive Ethernet: Evaluating & Optimizing Fast Inference Techniques for Deployment on Low-Cost Platform

2025-07-01

AI Summary

To address the challenge of real-time intrusion detection against traffic injection attacks in automotive Ethernet, this paper proposes a lightweight deep learning–based intrusion detection system (IDS) tailored for resource-constrained edge devices. To overcome the trade-off between inference latency and detection accuracy on low-end platforms (e.g., Raspberry Pi 4), we jointly apply knowledge distillation and structured pruning to design a compact detection model with significantly reduced parameters and computational cost, while optimizing the end-to-end inference pipeline. Experimental evaluation demonstrates that the system achieves a minimum per-sample inference latency of 727 µs on Raspberry Pi 4, with an AUC-ROC of 0.9890—substantially outperforming existing embedded IDS solutions. To the best of our knowledge, this is the first work to empirically validate the feasibility and robustness of a lightweight deep learning IDS in real-world automotive Ethernet edge environments under microsecond-level response constraints.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Large-scale security measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Modern vehicles are increasingly connected, and in this context, automotive Ethernet is one of the technologies that promise to provide the necessary infrastructure for intra-vehicle communication. However, these systems are subject to attacks that can compromise safety, including flow injection attacks. Deep Learning-based Intrusion Detection Systems (IDS) are often designed to combat this problem, but they require expensive hardware to run in real time. In this work, we propose to evaluate and apply fast neural network inference techniques like Distilling and Prunning for deploying IDS models on low-cost platforms in real time. The results show that these techniques can achieve intrusion detection times of up to 727 μs using a Raspberry Pi 4, with AUCROC values of 0.9890.

Problem

• Detect intrusions in automotive Ethernet using deep learning
• Optimize fast inference for low-cost hardware deployment
• Achieve real-time performance with minimal detection latency

Innovation

• Deep Learning-based Intrusion Detection for Automotive Ethernet
• Fast neural network inference techniques like Distilling and Pruning
• Deploying IDS models on low-cost platforms in real time