PEMANTAUAN POLA AKTIVITAS KERJA PADA SISTEM PEMANTAUAN KEBUGARAN PEKERJA DI INDUSTRI BERISIKO TINGGI

Ihsan Nur Hakim, Ir. Sentagi Sesotya Utami, S.T., M.Sc., Ph.D., IPU.; Ir. Nopriadi, S.T., M.Sc., Ph.D., IPM.

2025 | Tesis | MAGISTER TEKNIK FISIKA

Pemantauan pola aktivitas pekerja di industri berisiko tinggi telah dilakukan dengan Wearable Sensors (WS) untuk melengkapi proses Fit To Work (FTW) guna mengurangi kecelakaan kerja. Pekerja yang menggunakan WS adalah operator derek Ship To Shore (STS), Safety and Security Officer (SSO) yang bertugas patroli, dan Komandan Regu (DANRU) yang mengkoordinasikan SSO.

Data biomarker digital WS penelitian ini terdiri dari suhu tubuh, saturasi oksigen (SpO2), heart rate (HR), tekanan darah (TD) sistole dan diastole, heart rate variability (HRV), tingkat stres, dan dilengkapi ?terai WS dengan periode pencuplikan 10 menit. Sifat data dikaji dengan perhitungan statistik total dan agregat 1 jam beserta visualisasi dengan diagram violin, serta bootstrapping untuk menentukan perbedaan antargrup secara kuantitatif. Statistik total meliputi cacah, median, rerata, dan deviasi standar. Statistik agregat meliputi %kemunculan data, median, rerata, dan deviasi standar. Hasil bootstrapping berupa confidence interval (CI) dikonversi ke p-Value terkoreksi.

Data mentah pasca preprocessing dikaji statistik totalnya. Data mentah mengandung data aktif (saat WS dikenakan) dan data tak aktif (saat WS tak dikenakan). Bootstrapping statistik agregat digunakan untuk mengetahui variabel pembeda dari kedua grup tersebut. Ditemukan bahwa %kemunculan HR dan nilai rerata maupun median suhu adalah variabel pembeda kedua grup. Filter diterapkan pada data mentah dengan kriteria %kemunculan HR >= 50?n suhu dari 35 °C hingga 40 °C. Data terfilter dikaji statistik totalnya. Bootstrapping statistik agregat dari ketiga jenis pekerja diterapkan untuk mengetahui variabel pembeda antarjenis pekerja. Ditemukan bahwa variabel-variabel tersebut adalah nilai suhu, TD diastole, HRV, tingkat stres, HR, dan  ?terai, %kemunculan TD sistole, diastole, HRV, tingkat stres, HR, dan SpO2, serta deviasi standar dari HR, HRV, dan suhu. Variabel-variabel tersebut terindikasi berkaitan dengan pola aktivitas pekerja yang meliputi ketertiban penggunaan WS, gambaran umum posisi dan lingkungan di mana pekerja berada, keaktifan gerakan badan dan tangan, dan perubahan beban kerja.

Worker activity pattern monitoring in high-risk industry using wearable sensors (WS) has been implemented for complementing the Fit To Work (FTW) procedure to reduce accidents. The involved workers were Ship To Shore (STS) crane operators, Safety Security Officers (SSO) whose tasks were patrolling, and Squad Commander (DANRU) whose tasks were coordinating with SSO.

Digital biomarkers provided by the WS include body temperature, oxygen saturation (SpO2), heart rate (HR), systolic and diastolic blood pressure (SBP and DBP), heart rate variability (HRV), stress level, and WS battery percentage, with 10 minutes sampling period. The data characteristics were analyzed through total and hourly aggregate statistics with violin plots, as well as bootstrapping for quantitative group differences. The total statistics included count, median, mean, and standard deviation, while the aggregate statistics involved data occurrence percentage, median, mean, and standard deviation. The bootstrapping produced confidence intervals (CI), which then converted to adjusted p-values.

Raw preprocessed data was analyzed for the total statistics. The raw data consisted of active (worn WS) and inactive (unworn WS) data. The aggregate statistics bootstrapping revealed distinguishing variables between these data were HR occurrence percentage and mean or median temperature. Filtering was applied with criteria of HR occurrence ? 50% and temperature between 35 °C and 40 °C. The filtered data was analyzed for the total statistics. The aggregate statistics bootstrapping across worker types identified that the distinguishing variables are values of temperature, DBP, HRV, stress level, HR, WS battery percentage, and occurrence percentages of SBP and DBP, HRV, stress level, HR, and SpO2. Standard deviations of HR, HRV, and temperature also differed. These variables indicated relations to activity patterns, including WS usage compliance, positional and environmental contexts, body and hand movement, and workload changes.

Kata Kunci : Pemantauan, Pola Aktivitas, Wearable Sensors, Biomarker Digital, Industri Berisiko Tinggi / Monitoring, Activity Patterns, Wearable Sensors, Digital Biomarkers, High-Risk Industries