營建業管理人員工作疲勞監測系統之研究

Abstract

由於營建產業相對於其他行業的工作環境更為艱苦，存在較高的不確定性和風險因素，導致職業安全事故的發生率及死亡千人率遠高於其他行業。而造成工安事故有多種原因，包括人員、機具、工法及氣候等，其中異常的疲勞和過大的心理壓力是主要原因之一，因此除了探討工安意外的性質及工作環境的危險因子外，工人活動的實際生理狀況也是一個需要被關注的重點。此外，營建業的管理人員時常在龐大的工作負荷下到自身不熟悉的施工現場執行工作，這將提高職業災害發生的機率，由此可見若能預防工作疲勞所引發的不安全行為將有機會降低營建產業的職災發生率，為營建業人員增添一份保障。 過往研究雖有利用機器學習技術來預測疲勞及心臟病罹患機率等結果，但鮮少直接針對營建業管理人員的工作疲勞進行預測，亦或是用來建置模型的數據過於集中於某些特定人員，以致資料集不夠豐富，訓練之模型泛用性低。此外，隨著科技進步，生理感測裝置及傳輸技術突破以往的限制，使得本研究能夠以更輕便的裝置、更即時的傳輸技術來蒐集與應用高質量的生理資料。不僅如此，過往研究只能對疲勞狀況或生理指標做出預測的結果且相關的延伸內容只停留於理論階段，無法將其實際運用於現地的安全管理當中並預防風險發生，更重要的是營建業身為相對危險的產業卻沒有一個整體性的疲勞管理系統來確保相關人員的工作安全，使得職安人員不易從容地做到有效的負荷管理。 鑒於上述所言，本研究將使用生理感測手環搭配LoRa傳輸技術來蒐集營建業管理人員的心率資料，之後將使用資料進行工作疲勞預測並提出表現較佳之模型，除此之外將會設計個人的安全監測系統以觀察人員當前的生理狀態，最後將上述提及的兩部分交互應用以構築出一個綜合性的疲勞管理框架，如此使得職安人員得以遠端且即時對營建業之管理人員進行工作負荷管理，有效識別潛在的風險並提早針對可能的疲勞或壓力狀況採取適當的措施，避免其因身體疲勞或超出過往生理狀況而發生工作危害。

Due to the tough working environment in the construction industry compared to other industries, there is a higher level of uncertainty and risk factors, leading to a significantly higher incidence of occupational safety accidents and a higher mortality rate than in other industries. There are many causes of occupational accidents, including personnel, machinery, construction methods, and weather conditions. Among these, abnormal fatigue and excessive psychological stress are one of the main reasons. Therefore, in addition to exploring the essence of occupational accidents and hazardous factors in the working environment, the actual physiological condition of workers is also a critical focus. Furthermore, management personnel in the construction industry often work on unfamiliar sites under heavy workloads, which increases the probability of occupational accidents. It is evident that prevention of unsafe behaviors caused by work fatigue could reduce the accident rate in the construction industry, providing extra protection for the workers. Although previous studies have used machine learning techniques to predict fatigue and the likelihood of heart disease, few have directly targeted the work fatigue of construction industry managers, or the data used to train models has been overly concentrated on specific individuals, resulting in a data set that is not diverse enough and models with low versatility. Additionally, advancements in technology have overcome previous limitations in physiological sensing devices and transmission technologies, allowing this study to collect and use high-quality physiological data with lighter devices and more immediate transmission technologies. Moreover, previous studies were only able to make predictions about fatigue status or physical signs and their related extended content remained at a theoretical stage, which could not be practically applied in on-site safety management and prevent risks. More importantly, the construction industry, being a relatively hazardous sector, lacked a comprehensive safety management system to ensure the work safety of relevant personnel, making it difficult for occupational safety personnel to manage workloads effectively. Given the above, this study will use physiological sensing wristbands combined with LoRa transmission technology to collect heart rate data from management personnel in the construction industry. The data will then be used to predict work fatigue and propose a better-performing model. In addition, a personal safety monitoring system will be designed to ensure the current physiological state of the personnel. Finally, the aforementioned two components will be interactively applied to formulate a comprehensive safety management framework. This will enable occupational safety personnel to remotely manage the workload of construction industry managers, effectively identify potential risks, and take immediate measures against possible fatigue or stress conditions to prevent work-related accidents due to physical fatigue or conditions exceeding previous physical states.