Design and Implementation of a Prototype Smart Door Lock Integrating Face Recognition and Smart Doorbell for Residential Access Control
This paper presents the design and implementation of a prototype smart door lock that integrates YOLOv8n-based face recognition, a smart doorbell, and remote access control via Blynk and Supabase, demonstrating high accuracy and reliability in authenticating residents and managing visitor access for residential security.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Security at home has long relied on a simple, physical exchange: a key fits a lock, or a code opens a gate. While effective, these methods carry inherent risks; keys can be lost or stolen, and codes can be forgotten or overheard. In recent years, the concept of the "smart home" has promised to replace these mechanical tools with digital intelligence, using sensors and internet connections to manage who enters a house. A more advanced step in this evolution involves biometric systems, which use unique physical traits like fingerprints or facial features to identify people. Unlike a key that can be copied, a face is difficult to duplicate, offering a layer of security that is both personal and hard to forge. However, making such a system work reliably outside of a controlled laboratory, on a device small enough to fit on a front door, presents a significant engineering challenge.
Researchers at Hasanuddin University in Indonesia have tackled this challenge by building and testing a prototype smart door lock that combines facial recognition with a digital doorbell. Their goal was to create a single system that could automatically let in residents, alert homeowners to visitors, and even flag suspicious behavior, all without requiring a human to be present at the door. The device they constructed sits on a Raspberry Pi, a small, affordable computer often used for educational projects, which acts as the brain of the operation. Attached to this computer are a camera, a motion sensor, a solenoid lock that physically secures the door, and a button that functions as a doorbell. The system is designed to work in two distinct modes: one for people the house already knows, and another for strangers.
When someone approaches the door, a motion sensor triggers the camera to snap a picture. If the system recognizes the face as belonging to a registered resident, it instantly unlocks the door. This process happens entirely on the device itself, without needing to send the image to a distant server first, which keeps the response fast and private. If the face does not match anyone in the system's memory, the door remains locked. Instead of simply locking out the stranger, the system switches to a smart doorbell mode. It displays a message on a small screen asking the visitor to press the button. Once pressed, the homeowner receives an immediate notification on their smartphone, along with a photo of the person at the door. The homeowner can then choose to unlock the door remotely or leave it locked.
The researchers also programmed the system to watch for unusual behavior. If a person stands in front of the door and does not press the button within twenty seconds, the system assumes this might be suspicious activity. In such cases, it sends an alert to the homeowner immediately, even if no button was pressed, allowing the resident to investigate a potential threat without waiting for a call. To ensure the system worked correctly, the team tested it with a group of people, both registered users and strangers, at various distances from the door. They found that the camera could detect a face with perfect reliability up to three meters away. When it came to actually identifying the person, the system was highly accurate, correctly recognizing registered users about 92 percent of the time. Crucially, it never mistakenly unlocked the door for an unregistered person, maintaining a perfect record of rejecting strangers.
The entire process, from the moment a person is detected to the door unlocking or a notification being sent, takes less than five seconds on average. The system successfully delivered alerts to the homeowner's phone in about 2.6 seconds, and when the homeowner pressed the unlock button on their phone, the door opened in less than one second. The study concludes that this combination of local artificial intelligence and internet connectivity offers a practical, low-cost solution for modern residential security. By integrating the ability to recognize residents, manage visitors, and monitor for threats into one compact unit, the prototype demonstrates that sophisticated access control can be brought to everyday homes without requiring complex infrastructure or expensive hardware.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.