@inproceedings{35406146a69b40c9a8a72d44713cc6cb,
title = "GOTCHA: Physical intrusion detection with active acoustic sensing using a smart speaker",
abstract = "Voice Assistants (VAs) in the form of smart speakers such as Amazon Alexa, Google Assistant or Apple Siri are now commonplace. Due to their ubiquitous presence they can be useful as home alarm systems. In this work we present the development and evaluation of a novel physical intrusion detection system GOTCHA based on human presence detection with active acoustic sensing. GOTCHA can execute on simple off-the-shelf smart speaker hardware. Periodic audible chirps are employed to gather data which are then processed by a deep autoencoder trained on the acoustic profile of the empty room. As our evaluation shows, GOTCHA achieves promising results of up to 99.2\% for the F1-score. Our experiments show that a person can be detected without fail while the system barely generates false alarms. GOTCHA is a viable alternative to passive sensing solutions for physical intrusion detection.",
keywords = "Active acoustic sensing, Anomaly detection, Autoencoder, Deep Learning, Machine learning, Physical intrusion detection",
author = "Pham, \{Hoang Quoc Viet\} and Nguyen, \{Hoang D.\} and Utz Roedig",
note = "{\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2025. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.; 29th Nordic Conference on Secure IT Systems, NordSec 2024 ; Conference date: 06-11-2024 Through 07-11-2024",
year = "2025",
month = jan,
day = "29",
doi = "10.1007/978-3-031-79007-2\_18",
language = "English",
isbn = "9783031790065",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "345--363",
editor = "\{Horn Iwaya\}, Leonardo and Liina Kamm and Leonardo Martucci and Tobias Pulls",
booktitle = "Secure IT Systems - 29th Nordic Conference, NordSec 2024, Proceedings",
address = "Germany",
}