Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards Detecting IoT Event Spoofing Attacks Using Time-Series Classification

Uzma Maroof, Gustavo Batista, Arash Shaghaghi, Sanjay Jha

TL;DR

This work tackles IoT event spoofing in home automation by leveraging time-series classification in a dissimilarity-space representation built on Dynamic Time Warping. By mapping multivariate sensor data to a fixed-dimensional space using distances to carefully chosen prototypes, the approach preserves temporal structure while requiring far smaller training sets than traditional statistical-feature methods. Evaluated on the real-world PEEVES dataset, the method achieves comparable detection performance with up to 500x less labeled data and attains 100% accuracy for at least one event in end-to-end testing, highlighting practical viability for IoT deployments. The study also explores extensive model selection and downsampling strategies, and the authors release their system publicly to encourage further research and real-world experimentation.

Abstract

Internet of Things (IoT) devices have grown in popularity since they can directly interact with the real world. Home automation systems automate these interactions. IoT events are crucial to these systems' decision-making but are often unreliable. Security vulnerabilities allow attackers to impersonate events. Using statistical machine learning, IoT event fingerprints from deployed sensors have been used to detect spoofed events. Multivariate temporal data from these sensors has structural and temporal properties that statistical machine learning cannot learn. These schemes' accuracy depends on the knowledge base; the larger, the more accurate. However, the lack of huge datasets with enough samples of each IoT event in the nascent field of IoT can be a bottleneck. In this work, we deployed advanced machine learning to detect event-spoofing assaults. The temporal nature of sensor data lets us discover important patterns with fewer events. Our rigorous investigation of a publicly available real-world dataset indicates that our time-series-based solution technique learns temporal features from sensor data faster than earlier work, even with a 100- or 500-fold smaller training sample, making it a realistic IoT solution.

Towards Detecting IoT Event Spoofing Attacks Using Time-Series Classification

TL;DR

This work tackles IoT event spoofing in home automation by leveraging time-series classification in a dissimilarity-space representation built on Dynamic Time Warping. By mapping multivariate sensor data to a fixed-dimensional space using distances to carefully chosen prototypes, the approach preserves temporal structure while requiring far smaller training sets than traditional statistical-feature methods. Evaluated on the real-world PEEVES dataset, the method achieves comparable detection performance with up to 500x less labeled data and attains 100% accuracy for at least one event in end-to-end testing, highlighting practical viability for IoT deployments. The study also explores extensive model selection and downsampling strategies, and the authors release their system publicly to encourage further research and real-world experimentation.

Abstract

Internet of Things (IoT) devices have grown in popularity since they can directly interact with the real world. Home automation systems automate these interactions. IoT events are crucial to these systems' decision-making but are often unreliable. Security vulnerabilities allow attackers to impersonate events. Using statistical machine learning, IoT event fingerprints from deployed sensors have been used to detect spoofed events. Multivariate temporal data from these sensors has structural and temporal properties that statistical machine learning cannot learn. These schemes' accuracy depends on the knowledge base; the larger, the more accurate. However, the lack of huge datasets with enough samples of each IoT event in the nascent field of IoT can be a bottleneck. In this work, we deployed advanced machine learning to detect event-spoofing assaults. The temporal nature of sensor data lets us discover important patterns with fewer events. Our rigorous investigation of a publicly available real-world dataset indicates that our time-series-based solution technique learns temporal features from sensor data faster than earlier work, even with a 100- or 500-fold smaller training sample, making it a realistic IoT solution.
Paper Structure (18 sections, 7 equations, 5 figures, 4 tables)

This paper contains 18 sections, 7 equations, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Challenges
  4. Time-series Classification(TSC) & Design decisions
  5. Time Series Classification methods
  6. Dissimilarity measures
  7. Distance features
  8. Dissimilarity space
  9. Proposed System
  10. Constructing Dissimilarity space
  11. Model selection
  12. Evaluation
  13. Experimental Setup
  14. Performance metrics
  15. Training set size
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Statistical feature such as average signal strength unable to recognise that Signal 3 is a shifted version of Signal 1.
  • Figure 2: Comparison of dissimilarity measures.
  • Figure 3: Dissimilarity space mapping for a univariate dataset.
  • Figure 4: System Architecture.
  • Figure 5: Model Selection.