Adaptive Multi-Sensor Intrusion Detection System with Baseline Learning Phase for Real-Time Remote Alerting
International Journal on Science and Technology · 23.04.2026 · 17 (2)
DOI: 10.71097/ijsat.v17.i2.10874 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Science and Technology, online erschienen am 23.04.2026. Verfasst von G Vishnu Datta und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Security and surveillance systems play a critical role in protecting residential, agricultural, and industrial environments from unauthorized intrusions. However, conventional motion-based security systems often suffer from high false alarm rates caused by environmental disturbances such as wind, small animals, or background vibrations. This paper presents the design and implementation of an Adaptive Multi-Sensor Intrusion Detection System with Baseline Learning Phase for Real-Time Remote Alerting, developed using a low-cost ESP32 microcontroller platform. The proposed system integrates four directional PIR motion sensors and a vibration sensor to monitor physical activity across a protected perimeter. A key innovation of this system is its Baseline Learning Phase, in which the system autonomously observes and characterizes the surrounding environment for approximately 60 seconds upon startup. During this phase, two vibration parameters pulse frequency and active vibration duration ratio are sampled and averaged to establish a reference profile representing normal environmental conditions specific to the deployment location. Following baseline establishment, the system enters active monitoring mode and continuously computes a Disturbance Index, a weighted metric that quantifies deviation of real-time vibration data from the learned baseline. This index is combined with input from the four PIR sensors to classify the environment into one of three states: normal condition, suspicious activity, or confirmed intrusion. Upon confirmed intrusion detection, the system transmits an instant alert message via the Telegram messaging platform over Wi-Fi, delivering the direction of detected motion and the computed disturbance index value to the user's mobile device in real time
Zitieren
APA
Datta, G. V., & Majeed, M. (2026). Adaptive Multi-Sensor Intrusion Detection System with Baseline Learning Phase for Real-Time Remote Alerting. International Journal on Science and Technology, 17(2). https://doi.org/10.71097/ijsat.v17.i2.10874
Vancouver
Datta GV, Majeed M. Adaptive Multi-Sensor Intrusion Detection System with Baseline Learning Phase for Real-Time Remote Alerting. International Journal on Science and Technology. 2026;17(2). doi:10.71097/ijsat.v17.i2.10874
BibTeX
@article{Datta2026,
author = {Datta, G Vishnu and Majeed, Mehak},
title = {Adaptive Multi-Sensor Intrusion Detection System with Baseline Learning Phase for Real-Time Remote Alerting},
journal = {International Journal on Science and Technology},
year = {2026},
volume = {17},
number = {2},
doi = {10.71097/ijsat.v17.i2.10874}
}
EUVIMED