A Study On The Control Strategy For Typical Cut-in Accident Scenario Based On Drivers Behavior

The automotive intelligent industry is developing rapidly year by year,providing more efficient and intelligent solutions for all aspects of challenges faced by automotive technology.In recent years,as one of the key technologies in the field of automobile safety,the active collision avoidance technology of automobiles has become the main effective means to improve and solve the problems of vehicle collision accidents.However,in the face of the complex and changeable real traffic environment,an automobile collision avoidance system that only includes longitudinal braking obviously cannot cover all dangerous accident scenarios.Therefore,active collision avoidance technology needs to be refined for multiple typical working conditions to deal with more dangers.Scenes.In the daily driving environment,rear-end collisions and bevel collisions are very likely to occur when the front vehicle cutting-in.The incidence of such conditions accounts for 8% of the total accidents,but there are few targeted related studies.Therefore,this paper will carry out research on the active collision avoidance technology under the dangerous cut-in condition of the leading vehicle.First of all,this article will rely on the accident cases counted by the US Highway Administration(NHTSA),and select two typical dangerous cut-in conditions as the traffic scence analysis object.The scene is restored in the driving simulator,and the distance value at the time of the previous car’s cut-in time is used as the criterion for the degree of danger.The distance value is 15 m,30m,and 45 m.In order to obtain driver collision behavior data,Volunteers are recruited to carry out driving simulation experiments under this working condition,and then conduct driver collision avoidance behavior analysis based on the results of volunteer experiments.Data analysis results show that in the volunteer experiment,82.5%of the drivers adopt longitudinal braking to avoid collisions,while about 17.5% of drivers adopt steering and lane-changing operations to avoid collisions.Immediately collect the key parameters of the vehicle’s braking deceleration,braking time TTC and THW value during the experiment,and build the subsequent collision avoidance strategy based on the reference data value.Secondly,consider that the active collision avoidance system is a supplementary study of the full-condition collision avoidance system.Therefore,this paper builds the front vehicle cut-in state recognition module as a prerequisite for intervention in the collision avoidance system.Immediately build a collision avoidance strategy based on(Simulink/stateflow)tools,and combine the(MATLAB/driving Scenario Designer)module to perform a closed-loop test of the strategy logic to initially verify the logical correctness of the strategy.In the aspect of steering and changing lanes,this paper adopts five-order multinomial planning lane change path with simple expression and higher calculation efficiency.Based on the principle of model predictive control,the vehicle lane change path is tracked to achieve controller control accuracy and control stability.Finally,a joint simulation platform is built based on(Prescan/Matlab/Simulink)and(Carsim),and a total of 6 groups of typical working conditions of the above two types are simulated and verified.The final simulation results show that the system successfully avoids collisions,the parking distance is in the range of 0 ～ 5m,and the parking distance is in the reasonable range compared with mainstream collision avoidance systems.In the low-speed45km/h scene,the three sets of experiments were all adopted longitudinal braking to avoid collisions.The 45 m and 30 m groups applied decelerations at about 0.35 g and 0.75 g,respectively,and the speed of the main vehicle was changed.More gentle.The 15 m group has a higher degree of danger and brakes with a deceleration of about 0.7 ～ 0.8g;the 30 m and 45 m group in the high-speed 80km/h scene adopts longitudinal braking when the braking distance is sufficient,and the 15 m group with a higher degree of danger is longitudinally braked Unable to meet the needs of collision avoidance,we adopted steering to change lanes to avoid collision.In the aspect of turning and changing lanes,the peak value of the lateral tracking error of the lane-changing path does not exceed 0.14 m,accounting for only 0.037%of the total lateral distance of the lane-changing,and the vehicle heading angle changes smoothly,that is,the vehicle control is in good condition.Compared with the mainstream collision avoidance model based on longitudinal braking,the collision avoidance system predicts the cutting-in state of the front car 0.5 to 0.9s earlier when the front vehicle cuts into the lane,and provides steering collision avoidance options when the braking cannot avoid the collision.Improve the success rate of collision avoidance.From the selection and analysis of collision avoidance mode,the system uses longitudinal braking as the main collision avoidance mode,and the steering and lane-changing collision avoidance command is triggered under certain circumstances.The above research results show that the system characteristics fit well with the driver’s collision avoidance operation in the volunteer experiment.