Accurate Reconstruction of Traffic Accident Based on Multiple Optimization Algorithms and Evaluation of Craniocerebral Injury Risk / 医用生物力学

ABSTRACT

Objective To investigate the effect of different optimization algorithms on accurate reconstruction of traffic accidents. Methods Non-dominated sorting genetic algorithm-II ( NSGA-II), neighborhood cultivation genetic algorithm (NCGA) and multi-objective particle swarm optimization (MOPSO) were used to optimize the multi-rigid body dynamic reconstruction of a real case. The effects of different optimization algorithms on convergence speed and optimal approximate solution were studied. The optimal initial impact parameters were simulated as boundary conditions of finite element method, and the simulated results were compared with the actual injuries. Results NCGA had a faster convergence speed and a better result in optimization process. The kinematic response of pedestrian vehicle collision reconstructed by the optimal approximate solution was consistent with the surveillance video. The prediction of craniocerebral injury was basically consistent with the cadaver examination. Conclusions The combination of optimization algorithm, rigid multibody and finite element method can complete the accurate reconstruction of traffic accidents and reduce the influence of human factors.