Pengembangan Sistem Voice Assistant Berbasis Llm Dalam Environment Home Assistant(Studi Kasus: Perintah Suara Untuk Ruang Kelas Departemen Teknologi Informasi)

Yudha, Rendy Anfi (2025) Pengembangan Sistem Voice Assistant Berbasis Llm Dalam Environment Home Assistant(Studi Kasus: Perintah Suara Untuk Ruang Kelas Departemen Teknologi Informasi). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027211006_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (31MB) | Request a copy

Abstract

Penelitian ini membahas perancangan dan pengembangan sistem voice assistant berbasis Internet of Things (IoT) yang diintegrasikan dengan platform Home Assistant untuk mendukung implementasi Smart Classroom. Sistem ini dirancang agar guru dapat memberikan perintah suara guna mengontrol perangkat pintar di ruang kelas, seperti lampu, proyektor, dan AC. Untuk mencapai akurasi pengenalan suara yang andal, penelitian ini menggunakan kombinasi mikrokontroler ESP32-S3 dengan mikrofon I2S (INMP441), amplifier I2S (MAX98357A), dan speaker eksternal. Pengolahan suara dilakukan secara lokal dengan memanfaatkan fitur PSRAM pada ESP32-S3 untuk mendukung proses buffering dan pemrosesan sinyal secara real-time. Sistem ini menggunakan protokol Wyoming untuk menghubungkan komponen STT (Whisper), TTS (Piper), serta wake word detector (openWakeword dan microWakeword) dalam satu pipeline lokal. Hasil pengujian menunjukkan bahwa sistem mampu mengenali perintah dengan baik, memberikan respons yang sesuai, serta merespons wake word secara akurat hingga jarak 5 meter. Dengan dukungan pengolahan sinyal lanjutan dan penempatan perangkat yang optimal, sistem ini berpotensi untuk diimplementasikan sebagai solusi pengendalian perangkat pintar di ruang kelas secara efisien, responsif, dan tanpa kontak fisik.
=====================================================================================================================================
This study presents the design and development of a voice assistant system based on the Internet of Things (IoT), integrated with the Home Assistant platform to support Smart Classroom implementation. The system is designed to enable teachers to issue voice commands to control smart devices in the classroom, such as lights, projectors, and air conditioners. To ensure reliable voice recognition, the system utilizes a combination of an ESP32-S3 microcontroller, I2S microphone (INMP441), I2S amplifier (MAX98357A), and an external speaker. Voice processing is performed locally, leveraging the ESP32-S3’s PSRAM to support real-time buffering and signal processing. The system employs the Wyoming protocol to connect STT (Whisper), TTS (Piper), and wakeword detection components (openWakeword and microWakeword) within a single local pipeline. Testing results show that the system can accurately recognize commands, respond appropriately, and detect wakewords reliably at distances of up to 5 meters. With enhanced signal processing and strategic device placement, this system demonstrates strong potential as an efficient, responsive, and hands-free solution for managing smart classroom devices.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	IoT, Wyoming Protocol, Home Assistant, Smart Classroom.
Subjects:	T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Rendy Anfi Yudha
Date Deposited:	01 Aug 2025 02:40
Last Modified:	01 Aug 2025 02:40
URI:	http://repository.its.ac.id/id/eprint/124896

Actions (login required)

View Item