关键词: Kriging近似模型, 多目标优化, 可逆预紧式安全带, 约束系统, 自动紧急制动

Abstract: For the optimizations of combined effort between reversible pretension seatbelt (RPS) and airbag in reversible restraint systems(RRS). A simulation model of driver-side restraint systems from certain sedan is developed. And the model is validated about the dynamic response of occupant with regard to autonomous emergency braking (AEB) during pre-crash phase, as well as the injury outcomes within in-crash phase. On the basis of six key parameters of RRS are described as input variables, as well as four injury outcomes of head, thorax, neck, femur are described as objects. A comprehensive method involving design of experiments (DOE), Kriging model, and multi-object genetic algorithm II (MOGA II) is conducted to perform a multi-object optimization, and weighted injury criterion (WIC) is used to determine the optimum solution. The result has shown that the optimum solution of RRS improved occupant injury outcomes in both two frontal impact conditions compared to pre-optimized solution. In the condition of impact velocity is 56 km/h, optimum solution of RRS decreases the injury outcomes of thorax, neck and femur by the percent of 6.61%, 28.99%, 16.12% respectively, as well as 4.99% to WIC, but the injury outcome of head remains the same nearly. Nevertheless, the optimum solution of RRS provided better protection for occupant in the condition of impact velocity increased to 64 km/h, the decreased percent are 26.19%, 33.21%, 20.49%, 6.11% to head, thorax, neck and femur respectively, and 28.01% to WIC.

Key words: autonomous emergency braking, Kriging model, multi-object optimization, restraint systems, reversible pretension seatbelt