Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Efficient Adversarial Detection Frameworks for Vehicle-to-Microgrid Services in Edge Computing

Ahmed Omara, Burak Kantarci

TL;DR

This work tackles deploying adversarial detection for Vehicle-to-Microgrid (V2M) edge devices under tight resource constraints. It introduces an end-to-end framework that combines Neural Architecture Search (NAS) with model compression (pruning, quantization, and projection) to produce compact detectors without sacrificing robustness against FGSM, BIM, C&W, and CGAN attacks. The approach achieves substantial efficiency gains—memory reduced from 20 MB to 1.3 MB, inference time from 3.2 s to 0.9 s, and GPU utilization from 5% to 2.68%—while maintaining high detection accuracy (~92.7%). This enables real-time, privacy-preserving adversarial monitoring in constrained smart-grid edge environments and provides a practical path toward deployment.

Abstract

As Artificial Intelligence (AI) becomes increasingly integrated into microgrid control systems, the risk of malicious actors exploiting vulnerabilities in Machine Learning (ML) algorithms to disrupt power generation and distribution grows. Detection models to identify adversarial attacks need to meet the constraints of edge environments, where computational power and memory are often limited. To address this issue, we propose a novel strategy that optimizes detection models for Vehicle-to-Microgrid (V2M) edge environments without compromising performance against inference and evasion attacks. Our approach integrates model design and compression into a unified process and results in a highly compact detection model that maintains high accuracy. We evaluated our method against four benchmark evasion attacks-Fast Gradient Sign Method (FGSM), Basic Iterative Method (BIM), Carlini & Wagner method (C&W) and Conditional Generative Adversarial Network (CGAN) method-and two knowledge-based attacks, white-box and gray-box. Our optimized model reduces memory usage from 20MB to 1.3MB, inference time from 3.2 seconds to 0.9 seconds, and GPU utilization from 5% to 2.68%.

Efficient Adversarial Detection Frameworks for Vehicle-to-Microgrid Services in Edge Computing

TL;DR

This work tackles deploying adversarial detection for Vehicle-to-Microgrid (V2M) edge devices under tight resource constraints. It introduces an end-to-end framework that combines Neural Architecture Search (NAS) with model compression (pruning, quantization, and projection) to produce compact detectors without sacrificing robustness against FGSM, BIM, C&W, and CGAN attacks. The approach achieves substantial efficiency gains—memory reduced from 20 MB to 1.3 MB, inference time from 3.2 s to 0.9 s, and GPU utilization from 5% to 2.68%—while maintaining high detection accuracy (~92.7%). This enables real-time, privacy-preserving adversarial monitoring in constrained smart-grid edge environments and provides a practical path toward deployment.

Abstract

As Artificial Intelligence (AI) becomes increasingly integrated into microgrid control systems, the risk of malicious actors exploiting vulnerabilities in Machine Learning (ML) algorithms to disrupt power generation and distribution grows. Detection models to identify adversarial attacks need to meet the constraints of edge environments, where computational power and memory are often limited. To address this issue, we propose a novel strategy that optimizes detection models for Vehicle-to-Microgrid (V2M) edge environments without compromising performance against inference and evasion attacks. Our approach integrates model design and compression into a unified process and results in a highly compact detection model that maintains high accuracy. We evaluated our method against four benchmark evasion attacks-Fast Gradient Sign Method (FGSM), Basic Iterative Method (BIM), Carlini & Wagner method (C&W) and Conditional Generative Adversarial Network (CGAN) method-and two knowledge-based attacks, white-box and gray-box. Our optimized model reduces memory usage from 20MB to 1.3MB, inference time from 3.2 seconds to 0.9 seconds, and GPU utilization from 5% to 2.68%.

Paper Structure

This paper contains 7 sections, 1 equation, 3 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Proposed Framework
  4. Performance Evaluation
  5. Simulation settings
  6. Numerical Results
  7. Conclusion

Figures (3)

  • Figure 1: AI deployment pipeline
  • Figure 2: CNN detection model performance before compression
  • Figure 3: CNN detection model performance after compression