Sistem Deteksi Penghuni Ruangan Untuk Pengendalian Air Conditioner Menggunakan Metode YOLO

Sitanggang, Mathew HF (2026) Sistem Deteksi Penghuni Ruangan Untuk Pengendalian Air Conditioner Menggunakan Metode YOLO. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221003-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Penggunaan Air Conditioner (AC) yang masih dioperasikan secara manual sering kali menyebabkan pemborosan energi listrik karena AC tetap menyala meskipun ruangan tidak digunakan. Penelitian ini bertujuan untuk merancang dan mengimplementasikan sistem deteksi penghuni ruangan menggunakan metode You Only Look Once (YOLO) sebagai dasar pengendalian Air Conditioner secara otomatis. Sistem yang dikembangkan memanfaatkan kamera CCTV Reolink E1 Pro sebagai media akuisisi citra dan NVIDIA Jetson Orin Nano sebagai perangkat pemrosesan yang menjalankan model YOLOv8 untuk mendeteksi keberadaan penghuni secara real-time. Hasil deteksi dikirimkan ke ESP32 melalui jaringan Wi-Fi menggunakan protokol HTTP untuk mengendalikan AC menggunakan infrared transmitter. Sistem juga dilengkapi dengan sensor DHT22 untuk memantau suhu ruangan, LCD I2C sebagai media tampilan lokal, dan dashboard berbasis web untuk memantau kondisi sistem secara real-time. Pengujian dilakukan terhadap performa deteksi penghuni, komunikasi antara Jetson Orin Nano dan ESP32, pengiriman sinyal inframerah, logika pengendalian AC, serta konsumsi energi listrik sebelum dan sesudah penerapan sistem. Hasil pengujian menunjukkan bahwa seluruh fungsi sistem berjalan sesuai dengan rancangan, dengan komunikasi antara Jetson Orin Nano dan ESP32 memiliki waktu respons rata-rata sebesar 55,63 ms sehingga mampu mendukung pengendalian AC secara cepat dan andal. Analisis konsumsi energi menunjukkan bahwa penerapan sistem berhasil menurunkan konsumsi energi listrik dari 114,22 kWh menjadi 95,71 kWh, sehingga diperoleh penghematan energi sebesar 18,51 kWh atau sebesar 16,20% dibandingkan pengoperasian secara konvensional. Hasil penelitian menunjukkan bahwa sistem deteksi penghuni berbasis YOLO mampu mengendalikan AC secara otomatis sesuai dengan keberadaan penghuni, sehingga dapat meningkatkan efisiensi penggunaan energi listrik tanpa mengurangi kenyamanan pengguna.
=====================================================================================================================================
The continuous operation of an Air Conditioner (AC) without considering room occupancy often results in unnecessary electrical energy consumption because the AC remains active even when the room is unoccupied. This study aims to design and implement a room occupant detection system using the You Only Look Once (YOLO) method as the basis for automatic Air Conditioner control. The proposed system employs a Reolink E1 Pro CCTV camera for image acquisition and an NVIDIA Jetson Orin Nano as the processing unit running the YOLOv8 model to detect room occupants in real time. The detection results are transmitted to an ESP32 via a Wi-Fi network using the HTTP protocol, enabling the ESP32 to control the Air Conditioner through an infrared transmitter. The system is also equipped with a DHT22 temperature sensor, an I2C LCD for local display, and a web-based dashboard for real-time system monitoring. Performance evaluation was conducted on occupant detection, communication between the Jetson Orin Nano and ESP32, infrared signal transmission, AC control logic, and electrical energy consumption before and after system implementation. The experimental results demonstrate that all system components operated according to the proposed design, while the communication between the Jetson Orin Nano and ESP32 achieved an average response time of 55.63 ms, enabling fast and reliable real-time AC control. Furthermore, the energy consumption analysis showed that the proposed system reduced electrical energy consumption from 114.22 kWh to 95.71 kWh, resulting in an energy saving of 18.51 kWh or 16.20% compared with conventional AC operation. These findings indicate that the proposed YOLO-based room occupant detection system can automatically control the Air Conditioner according to room occupancy, thereby improving energy efficiency without compromising occupant comfort.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Air Conditioner, efisiensi energi, deteksi penghuni, ESP32, NVIDIA Jetson Orin Nano, Air conditioner, energy efficiency, occupancy detection, ESP32, NVIDIA Jetson Orin Nano.
Subjects:	T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Mathew Hf Sitanggang
Date Deposited:	07 Aug 2026 07:42
Last Modified:	07 Aug 2026 07:42
URI:	http://repository.its.ac.id/id/eprint/143315

Actions (login required)

View Item