Syaiful Alam, Ageng Sadnowo Repelianto, Sri Purwiyanti, A Sedjahtera, Ahmad Saudi Samosir

Rancang Bangun Kunci Elektronik Tervalidasi dengan Suara dan Foto Pengakses melalui Aplikasi Telegram

Introduction

Rancang bangun kunci elektronik tervalidasi dengan suara dan foto pengakses melalui aplikasi telegram. Rancang bangun kunci elektronik aman: validasi suara & foto via Telegram. Voice recognition dan ESP32 Cam pastikan keamanan pintu optimal. Akurasi pengenalan suara 95-100% (10-80cm).

Abstract

Penelitian ini bertujuan untuk membangun prototype sistem kunci elektronik menggunakan modul voice recognition v3 dengan sistem validasi oleh pemilik otoritas melalui foto yang dikirim via aplikasi telegram. Ada dua kunci (solenoid door lock) yang perlu terbuka agar dapat membuka pintu. Kunci diakses dengan pertama tama memvalidasi suara dengan modul voice recognition. Validasi ini berdasarkan kata kunci yang sesuai. Kesesuaian ini bukan hanya pada kata, tapi juga intonasi dan kejernihan suara. Jika sesuai, maka salah satu kunci akan terbuka. Selain itu, bersamaan dengan validasi suara itu, foto pengakses akan diambil oleh modul ESP32 Cam dan dikirim melalui aplikasi telegram ke pemilik otoritas untuk memperoleh validasi kedua. Jika diizinkan, maka pemilik otoritas akan mengirim sinyal kendali melalui aplikasi itu untuk membuka kunci yang kedua. Pemilik otoritas dapat mengendalikan kunci baik untuk membuka maupun mengunci pintu. Dari hasil pengujian diperoleh bahwa sistem pengaman pintu telah berhasil bekerja dengan baik. Modul voice recognition v3 dapat beroperasi baik di jarak 10 - 80 cm yang memiliki tingkat pengenalan suara dengan satu suku kata sebesar 95% hingga 100%. Selain itu, suara yang diterima perlu sesuai dengan intonasi dan warna suara yang telah direkam ke modul itu terlebih dulu. Sistem pengambilan dan pengiriman foto oleh ESP32-Cam juga bekerja dengan baik, demikian pula pengendalian kunci baik membuka maupun mengunci melalui aplikasi telegram.

Review

The submitted paper presents a prototype of an electronic lock system that integrates multi-modal validation for enhanced security. Titled "Rancang Bangun Kunci Elektronik Tervalidasi dengan Suara dan Foto Pengakses melalui Aplikasi Telegram" (Design and Build of Validated Electronic Lock with Voice and Photo Accessors via Telegram Application), the research introduces an innovative approach to access control. The core contribution lies in its two-stage validation process: initial voice recognition using a V3 module, followed by a real-time photo capture and remote authorization by an authority figure via the Telegram application. This dual-key mechanism, requiring both local voice authentication and external human approval, positions the system as a robust solution for controlled access environments. Methodologically, the system effectively leverages readily available and cost-effective components, namely the voice recognition V3 module and an ESP32 Cam, integrated with the ubiquitous Telegram platform. The experimental results, as presented, indicate a successful implementation and satisfactory performance of individual modules. Specifically, the voice recognition module demonstrated high accuracy (95-100%) for single-syllable recognition within a 10-80 cm range, emphasizing the importance of matching intonation and voice tone. The seamless operation of the ESP32 Cam for image capture and transmission, alongside the Telegram application's utility for remote control and validation, are also commendable. The design choice to have a human authority validate the access in real-time adds a critical layer of security, mitigating potential bypass attempts inherent in purely automated systems. While the prototype demonstrates promising functionality, the abstract raises several points for deeper consideration and future development. The reported voice recognition accuracy is impressive for single syllables, but real-world scenarios typically involve multi-syllable words or phrases, which might present a greater challenge for the V3 module. Further research should explore the system's performance under varying environmental conditions, such as background noise, which can significantly impact voice recognition reliability. Additionally, the abstract briefly touches upon the "owner authority" but does not elaborate on contingency plans if this authority is unavailable or unresponsive, which is a crucial aspect for a practical security system. Future iterations could also benefit from an analysis of response times for the remote validation process, comprehensive security audits of the Telegram integration, and robust error handling mechanisms, particularly regarding network connectivity failures or false positives/negatives in both voice and photo recognition. These enhancements would further solidify the system's practicality and resilience in real-world deployments.

Full Text

Comments

You need to be logged in to post a comment.