Automatic Door Lock Based on Knock Pattern and Face Detection

Abstract

Residential burglaries are prevalent offenses, frequently occurring during the absence of homeowners. These offenses
generally entail breaching doors or windows. An innovative home security system is important to resolve this issue. This research
seeks to create an automated door locking mechanism utilizing knock patterns and facial recognition to improve residential security.
The system incorporates a piezoelectric sensor for detecting knock patterns and an ESP32-Cam for facial recognition. The study
methodology entails the development and evaluation of a system that integrates two primary components, guaranteeing that the
door unlocks solely upon the recognition of both an accurate knock pattern and a registered facial image. The system's accuracy
was assessed under varying lighting situations and distances to determine its efficacy. The results indicate that the face detection
system operates effectively under optimal lighting settings, and the knock pattern system activates the door lock mechanism when
the knock intervals correspond to the pre-registered pattern. Nonetheless, the system encounters difficulties in recognizing faces in
low-light conditions. The door lock is secure, as it will only unlock when both the appropriate facial recognition and knock pattern
criteria are satisfied, thus improving security relative to conventional locks. This dual-layer security strategy mitigates the dangers
inherent in traditional systems, such as keys or PINs, which are susceptible to theft or circumvention. The proposed technology
significantly enhances home security, presenting a more secure and user-friendly alternative to current options. Future
enhancements may concentrate on augmenting the precision of facial detection in low-light environments and refining the system
for wider practical applications.