DEVELOPMENT OF PERFORMANCE PREDICTION MODEL AS FATIGUE FUNCTION ON DRIVING TRAIN SIMULATOR ACTIVITIES

• Main Author: VALENTINO NIM: 23417011, FANDY
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/27140
• Institution: Institut Teknologi Bandung

Human aspect is still one of the dominant aspects of the cause of high rail transportation accidents in Indonesia. In the past seven years there have been as many as 35 accidents with 33% of them caused by humans associated with fatigue and sleepiness on duty. Factors that affect this level of fatigue include the quantity of the previous sleep and the monotonous task. Sleep deprivation and monotonous conditions can cause the level of vigilance in work to decrease and the risk of accidents increases. So far the measurement of fatigue and sleepiness in machinists is still limited. Measuring the level of readiness of work in machinists before work only includes measurement of blood pressure, body temperature, and pulse rate. This is not enough to be an indicator of whether a machinist can be considered worthy of duty. This study aims to develop a model that can predict performance in rail driving activity using fatigue gauges and validate it. <br /> <br /> To achieve this goal, an experiment was conducted using a four-hour rail simulator to represent the driving activity. To get data that varies, participants first limited sleeping time to two, four and eight hours. In addition, the characteristics of the driving tasks performed are divided into monotonous and dynamic conditions. Measurements were performed using Sustained Attention Test (SAT) and Psychomotor Vigilance Task (PVT) to assess the level of alertness and reaction velocity in response to a set of Swedish Occupational Fatigue Index (SOFI) stimuli to measure fatigue level, Karolinska Sleepiness Scale (KSS) sleepiness level and Visual Analogue Scale (VAS) to measure how much desire to rest. The parameters of these various measurements then become independent variables for building linear regression models to predict performance during railway driving. <br /> <br /> Based on the research that has been done, it can be concluded that model can be developed to predict the performance with equation <br /> <br /> With LE, PD, MSAT and LPVT respectively states Lack of Energy, Physical Discomfort, Miss-Sustained Attention Test and Lapse-Psychomotor Vigilance Task. The above equation has a coefficient of determination is 25.70%.