Sistem Pelacakan Posisi Petugas Kebersihan Menggunakan ESP32 dan BLE

Abstract

Pemantauan lokasi petugas kebersihan di dalam gedung merupakan salah satu upaya meningkatkan efektivitas pengelolaan fasilitas, namun sistem Indoor Positioning System (IPS) berbasis Bluetooth Low Energy (BLE) dan ESP32 yang ada masih menghadapi kendala seperti kalibrasi RSSI yang dilakukan secara manual, pencatatan data yang terus-menerus sehingga membebani basis data, serta pertumbuhan data historis yang menyebabkan pembengkakan ruang penyimpanan. Penelitian ini bertujuan mengembangkan sistem pelacakan posisi petugas kebersihan berbasis ESP32 dan BLE dengan menambahkan fitur kalibrasi RSSI melalui antarmuka web, pemantauan sinyal dan deteksi anomali/interferensi secara otomatis, manajemen pendaftaran perangkat, visualisasi dwelling time melalui heatmap durasi, serta Auto-Data Pruning untuk menjaga efisiensi penyimpanan basis data. Sistem dikembangkan menggunakan ESP32 sebagai pemindai sinyal BLE, Flask sebagai backend, React Vite sebagai frontend, dan PostgreSQL sebagai basis data, dengan estimasi posisi dihitung melalui metode trilaterasi berbasis RSSI. Pengujian fungsionalitas dilakukan melalui Black Box Testing dengan pendekatan Equivalence Partitioning serta evaluasi akurasi posisi menggunakan Mean Absolute Error (MAE). Hasil pengujian menunjukkan sistem kalibrasi berbasis web, deteksi anomali sinyal, manajemen perangkat, dan visualisasi dwelling time berfungsi sesuai rancangan, dengan MAE estimasi posisi sebesar 0,98 meter, sedangkan mekanisme Auto-Data Pruning masih memerlukan perbaikan lebih lanjut.

Monitoring the locations of cleaning staff inside buildings is one way to improve the effectiveness of facility management; however, existing Indoor Positioning Systems (IPS) based on Bluetooth Low Energy (BLE) and ESP32 still face challenges such as manual RSSI calibration, continuous data logging that burdens the database, and the accumulation of historical data that leads to excessive storage space usage. This research aims to develop an ESP32- and BLE-based cleaning staff positioning tracking system by adding features such as RSSI calibration via a web interface, automatic signal monitoring and anomaly/interference detection, device registration management, dwelling time visualization through duration heatmaps, and Auto-Data Pruning to maintain database storage efficiency. The system was developed using ESP32 as the BLE signal scanner, Flask as the backend, React Vite as the frontend, and PostgreSQL as the database, with position estimates calculated using an RSSI-based trilateration method. Testing was conducted via Black Box Testing using the Equivalence Partitioning approach, and position accuracy was evaluated using Mean Absolute Error (MAE). Test results show that the web-based calibration system, signal anomaly detection, device management, and dwelling time visualization functions as designed, with a position estimation MAE of 0.98 meters, while the Auto-Data Pruning mechanism still requires improvement.