Effectiveness Evaluation of Biomathematical Model of Fatigue Based on Chinese Civil Aviation Crew Data

In order to develop a biomathematical model of fatigue suitable for the fatigue risk management of Chinese civil aviation crews, this paper first makes a series of improvements to the sleep algorithm of the classic three process model of alert (TPMA), and then tests the TPMA and the modified TPMA model (mTPMA) based on six groups of Chinese civil aviation crew data. Statistical analysis was used to compare the sleep/wake time difference and average sleep duration between the model output data and the data collected by the ACTIGRAPH, and the effectiveness of the mTPMA sleep prediction results was evaluated in combination with the success rate of sleep prediction. Linear regression analysis was used to compare the similarity of TPMA, mTPMA and FAID's output KSS curves to evaluate the effectiveness of mTPMA fatigue prediction. The data analysis results show that the success rate of mTPMA for sleep prediction (97.5 %) is significantly better than that of TPMA (93.2 %). mTPMA has good accuracy in the time difference of falling asleep, and there is no significant difference in the time difference of awakening and average sleep duration between mTPMA and ACTIGRAPH data. The goodness of KSS curve fitting of mTPMA and FAID is significantly improved. mTPMA has good sleep interval and fatigue level prediction ability, which can be realized for Chinese civil aviation crew.