應用穿戴裝置即時生理感測系統發展營建個人化工作負荷管理

Abstract

營建業職災率偏高，職災事件常造成嚴重工期延宕及賠償成本，許多研究指出過度疲勞為勞工發生意外的原因之一，但目前營建業安全管理尚缺乏導入即時工作負荷生理感測系統技術。本研究以光體積變化掃描圖技術(photoplethysmography, PPG)穿戴式心率感測手環，發展即時生理感測物聯網系統，以遠距即時蒐集勞工心率，目的為建立勞工場域休息心率(field resting heart rate, FRHR)個人基準心率值，並建立儲備心理百分比(the percentage of heart rate reserve, %HRR)工作負荷指標之安全管理區間。生理感測系統數據信度與效度驗證部分，在實驗室由5位受測者在設定運動模式下比對PPG與ECG感測心率，所選用PPG心率手環與獲美國FDA認證之胸貼式心電圖(electrocardiogram, ECG)心率感測驗證結果，平均絕對誤差百分比(mean absolute error percentages, MAPE)低於5%，相關係數r介於0.8128~0.9947，具有高度相關性，故本研究PPG心率手環具相當可靠度及準確度。在潛盾隧道工地場域，驗證數據覆蓋率(sensing data coverage rate) 81.35%，採取複合策略包括PPG手環心率訊號10次/秒以上之高頻率數據傳輸頻率、工區適當的BLE-Ethernet接收傳感器配置及5分鐘時間窗數據處理模式等，可建立生理感測系統效度。 因應工區場域實際施工作業之狀況，本研究提出適用於個人化FRHR之演算模式，經成對t 檢驗(paired t-test)及成對變數差異分析(paired variable difference analysis)，結果以分鐘計量平均值取值之最低單一值(WHRmin1)萃取模式，可獲得具穩定趨勢與時效優勢之休息心率，避免突發極值干擾；再者，藉歷史累積感測心率數據庫，至少5工作日逐步優化，可得個人FRHR之收斂值( < 3 bpm)，整體結果顯示FRHR演算模式可應用於儲備心率百分比工作負荷指標。由PPG心率感測系統產出潛盾隧道施工人員之個人工作負荷%HRR，勞工體能負荷程度%HRR與累積工時百分比關係呈現S曲線關係，其分布位置及反曲點具有工作負荷安全管理上意義。以感測系統歷史數據，累積多條正常工作負荷情況下S型曲線，所形成閉合之區間帶，可建立每一位勞工個人對該作業任務之體能負荷合理安全區間及邊界。此外，系統即時產出數據，可延伸應用計算每日勞工個人化之%HRR負荷之偏態、峰度係數及50%累計工時所對應之%HRR值(%HRR50)，用以顯示勞工個人在該工作日一半的施工活動時間所必須付出的體能負荷水準，評量工時調度及工作效率。 即時心率感測系統所建立之%HRR體力工作負荷指標，以個人化之相對性體能基準，利用大數據優化設定工作負荷分級判斷準則，減少監測系統過度工作負荷假訊號，有助於個人工作體能負荷監測精準管理。本研究所發展之PPG即時心率感測系統發展方式、個人場域休息心率基準FRHR演算模式，及%HRR相關指標，可供未來其他營建工程應用，增進營建業勞工安全管理輔助，促進勞工身心健康及安全管理，減少職災意外發生。

The occupational accident rate is relatively high in the construction industry. Major accidents seriously postponed construction periods and cost the compensation of victims and projects. Previous studies pointed excessive fatigue of workers is one of the reasons for the occupational accident occurring. However, the safety management in the construction industry still lacks to introduce the technologies for real-time physiological sensing workload systems currently. This present study adopted the wearable PPG-based heart rate sensing wristbands and developed a real-time physiological sensing IoT system for monitoring and collecting workers' heart rates consecutively. This study aimed to establish the personnel's physical baselines of field resting heart rate (FRHR) and apply it to the safety zone of workload with the percentage of heart rate reserve (%HRR), a relative physiological indicator for the workload. First, the PPG-based heart rate sensing devices and the physiological sensing system verified the reliability and validity of their data accuracy and collection coverage rate. The sensing heart rate of selected PPG wristbands was compared with the chest-worn electrocardiogram (ECG) certified by US FDA from 5 subjects in the set exercise mode in the laboratory. The verification results showed that the PPG heart rate wristbands used in this study were reliable and accurate with a high correlation with ECG, mean absolute error percentages (MAPE) < 5%, and the correlation coefficient r between 0.8128 ~ 0.9947. The physiological sensing system adopted multiple strategies to establish a sensing data coverage rate of 81.35%, including the high-frequency signals transmission of more than ten times per second, the appropriate layout of BLE-Ethernet Sensor-Hubs in the sensing zones, and data processing mode of a 5-minute time window in the experimental field of the shield tunnel worksites. For the characteristics of practical construction management on site, this study proposed the personalized FRHR algorithm model for HRR calculation. The results of paired t-test and paired variable difference analysis indicated that the data extraction mode of the lowest single value (WHRmin1) with the average minute measurement could acquire the resting heart rate. It came to the advantages of stable numerical value and time-effectiveness, which could avoid the interference of outliers. Additionally, optimizing the personal FRHR gradually for at least five working days from the historical database of sensing data could achieve their convergences of less than three bpm. The FRHR algorithm model was facilitated to individual workload indicator of %HRR rather than a single measurement value to determine the fitness baselines. The relationship between the %HRR levels of workers and the percentage of cumulative working hours has been plotted in an S-curve based on individual heart rate data from the PPG heart rate sensing system. The curve locations and their inflection points had significance in workload safety management. Multiple S-shaped curves formed a reasonable workload safety zone from historical %HRR data under normal workload conditions. It provided the boundary of safety management for alerting overloaded or abnormal physical demands of individual workers' physical workload for tasks. Additionally, the %HRR50 corresponding to 50% of the accumulated working hours indicated the physical demand level of tasks that individual workers need to pay for half of the working day, which could be a task assign management indicator based on physical fitness level for working hours scheduling and efficiency. This research developed a real-time heart rate sensing system. It established a %HRR physical workload index based on the personalized relative physical baselines, which could reduce false signals of the excessive workload of the monitoring system for precise management of individual workload. The installation criteria of the PPG real-time heart rate sensing system, the personnel FRHR baselines algorithm, and the % HRR-related indicators can be applied for other construction projects in the future. Improving safety management assistance at construction worksites and promoting workers' physical and mental health to reduce occupational accidents will be beneficial.