Anomalous IoT Behavior Detection by LSTM-Based Power Waveform Prediction

Ryusei Eda, Nozomu Togawa

2025

Abstract

Internet of Things (IoT) devices have very rapidly spread out in recent years. In IoT devices where applications run on operating system (OS), the power consumption of the OS and the power consumption of the applications overlap, resulting in complex power waveform. Previous methods need to explicitly extract the application power waveform from the multiple signal sources in the measured power waveform, which often fail to detect anomalous behaviors. In this paper, we propose a method to detect anomalous behaviors by using LSTM (Long Short Term Memory). The proposed method learns power waveform containing multiple signal sources and compares the predicted waveform and the actual one. Then, we can successfully detect anomalous behaviors, even though the measured power waveform is composed of multiple signal sources. Experimental results show that anomalous behavior can be successfully detected from an IoT device built with Raspberry Pi4.

Paper Citation

in Harvard Style

Eda R. and Togawa N. (2025). Anomalous IoT Behavior Detection by LSTM-Based Power Waveform Prediction. In Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS; ISBN 978-989-758-750-4, SciTePress, pages 345-352. DOI: 10.5220/0013368900003944

in Bibtex Style

@conference{iotbds25,
author={Ryusei Eda and Nozomu Togawa},
title={Anomalous IoT Behavior Detection by LSTM-Based Power Waveform Prediction},
booktitle={Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS},
year={2025},
pages={345-352},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013368900003944},
isbn={978-989-758-750-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 10th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS
TI - Anomalous IoT Behavior Detection by LSTM-Based Power Waveform Prediction
SN - 978-989-758-750-4
AU - Eda R.
AU - Togawa N.
PY - 2025
SP - 345
EP - 352
DO - 10.5220/0013368900003944
PB - SciTePress