Abstract: Child restraint systems designed to match child occupant only under one collision condition can’t provide effective protection for children under other collision condition. An adaptive child restraint system, which characteristic parameters are adjustable, is used to analyze the optimum parameters of restraint system for child occupants at different ages. The seat height, the cushion length and the force limiter value of the seat belt are chosen as optimization variables, and children’s comprehensive damage evaluation I and neck injury threshold Nkm are chosen as optimization objectives. The methods of approximate model of polynomial regression and non-dominated sorting genetic algorithm II are combined to conduct, multi-objective optimization for 3, 6, 10 years old child dummy respectively in both the front impact and rear impact according to GB 27887. The results show that the optimized parameters of restraint system can provide a better protection for children of the three age group in the two collision conditions. The correlation model between optimum parameters and child’s height and weight is fitted based on optimization design points of the three age groups. Therefore, good protection effect can be achieved to children with different height and weight.

Key words: adaptive child restraint system, different collision condition, multi-objective optimization, non-dominated sorting genetic algorithm II(NSGA-Ⅱ)