Real-Time Detection of Intruders Using an Acoustic Sensor and Internet-of-Things Computing.

Area of Science:

• Electrical Engineering
• Computer Science
• Security Systems

Background:

• Modern home automation systems increasingly integrate security features like cameras and radars.
• Existing systems may have limitations in detecting subtle or unexpected intrusions.
• There is a need for advanced, user-friendly security solutions leveraging emerging technologies.

Purpose of the Study:

• To propose and evaluate an innovative home security system utilizing acoustic signal analysis.
• To develop an Internet of Things (IoT)-based system for real-time burglar detection and notification.
• To implement and test an adaptive detection algorithm for enhanced security.

Main Methods:

• The proposed system employs an adaptive detection algorithm: "short-time-average through long-time-average" (STA-LTA).
• Implementation is achieved using an Internet of Things (IoT) device, specifically an Arduino Duo.
• The system analyzes acoustic signals from various sources like footsteps, falling objects, and speech for security purposes.

Main Results:

• The system demonstrated effectiveness in detecting acoustic events indicative of potential security breaches.
• Performance was evaluated using a set of 10 distinct acoustic signals, including background noise.
• The influence of various algorithm parameters on system performance was thoroughly investigated.

Conclusions:

• The developed IoT-based acoustic security system offers a novel approach to home and office protection.
• The adaptive STA-LTA algorithm shows promise for reliable intruder detection through sound analysis.
• The system provides a user-friendly and networked solution for enhanced security.