Study Of The Driver's Instinctive Response At The Collision Moment And The Influence On The Collision Safety Of The Lower Extremity

As a key part of the driver-vehicle-road closed-loop system,the researches about driver behavior have always been extensively focused.The driver's instinctive response is considered as the instinctive behavior when facing oncoming emergency and it has a great influence on the design of vehicle safety systems.In this thesis,the driver's instinctive response and its influence on the collision injury and the protection system were investigated.The main contents are listed as follow.Firstly,a simulated collision scene was constructed in a driving simulator,and 20 male volunteers and 20 female volunteers were recruited for the driving behavior test.The instinctive behavior,including the vehicle control parameters and the electromyogram of lower extremity,were recorded in different emergent level.The electromyogram signal was denoised using CIU united algorithm.The muscle activation moments that representing the instinctive response time and the muscle activation level that representing the musculoskeletal characteristics accompanning with the instinctive response posture were obtained.Secondly,the electromyogram was introduced into the investigation of driver response time,and it was divided into pre-motor time,muscle activation time,accelerator release time and movement time,then their influence by the factors were studied.The results showed that,the response time showed significant difference between different velocities,distancse between the host and the obstacle vehicle,the muscles and the genders.Tibia anterior is the first reacted muscle after the emergency occurs,so it can be used as the representation of the driver's brake intention.Muscle activation time is considered as the time that can be saved after the electromyogram was introduced,and the driver's brake intention is detected earlier.The driver's lower extremity posture at the collision moment was classified,combing the vehicle control parameter.A two stage posture classifier was constructed based on the generalized linear model,and the precise estimation of the brake posture at the collision moment was realized.The brake behavior in different emergent level was also analyzed.The result showed that the driver's right leg is possible to be on the brake pedal,in the air or on the accelerator,and it's more likely to be in the air or ever on the accelerator with high emergent level.Then,collision injury simulation was performed in different muscle activation state and brake posture for the lower extremity,and their influence on injury was studied.The injury in the accurate muscle activation state in typical condition was acquired,and compared tothose in inactivate,median and maximum state,and the meaning to describe the accurate muscle activation state was illustrated.Then,three brake postures,on the brake pedal,in the air and on the accelerator,were considered in the injury analysis,and the differences between two legs were also compared.The results showed that the injury of lower extremity increased with muscle activation state and the driver is not always in the median and maximum activation state.The injury was different when the lower extremtity was in different posture.The muscle activation state was affected by the emergent level,gender and posture.Therfore,it is important to precisely measure the muscle activation state at the collision moment,and to take the influence of emergent level,brake posture,muscle and gender into account,for better investigating the lower extremtity injury.Finally,the influence of the instinctive response on the injury was studied,combining the control behavior and the relationship between the brake behavior and the musculoskeletal characteristic at the collision moment.The injury parameters increased with emergent level no matter whether the knee aigbag was equipped or not.The influence of muscle activation showed important when the emergent level was comparatively small.The injury parameters of right leg were larger than that of left leg as for the influence of pedal.An injury model was constructed while considering the protection performance of the knee airbag.The ignition time,mass-velocity ratio and vent hole ratio were optimized based on a multi-object optimization algorithm of response surface method and NSGA-II,and the injury parameters decreased in different degree after optimization.The estimation model based on extreme learning machine was constructed for the three brake posture while considering the driver's instinctive response,to predict the key airbag parameters in actual situation.