Research On Test Scenarios And Evaluation Method Of Autonomous Emergency Braking System Based On Accident Data

The autonomous emergency braking system(AEB)plays a significant role in reducing traffic accidents and improving vehicle driving safety,however,testing and evaluating its safety and braking effectiveness are prerequisites for its widespread application.To test the performance of AEB,a scientific and comprehensive AEB testing and evaluation system needs to be constructed.AEB testing and evaluation mainly includes testing scenarios and comprehensive test evaluation.And the current AEB testing and evaluation faces issues such as simple testing scenarios,inadequate consideration of complex scenarios,and single evaluation indicators in existing evaluation methods,which make it difficult to objectively reflect the actual performance of the AEB system.Aiming at the above issues,this paper conducts in-depth research on the test scenarios and evaluation methods of autonomous emergency braking systems,in order to enrich the AEB test scenarios and objectively and comprehensively evaluate the actual performance of the AEB system.The main research contents are as follows:(1)Established a set of AEB testing scenarios.Typical AEB testing scenarios were extracted based on the collision accident data of the National Automobile Accident In-Depth Investigation System(NAIS).In order to make up for the shortcomings of dangerous testing scenarios in existing standards and regulations,started from the actual traffic road environment in China,screened the scene elements according to NAIS passenger car crash data,and designed the scene element extraction software.The commonalities of AEB testing scenarios in existing regulations were analyzed,and standard testing scenarios were extracted and combined with typical testing scenarios to form an AEB testing scenario set.(2)AEB comprehensive evaluation method with multiple evaluation indicators was proposed.Aiming at the problem of considering single evaluation indicators in existing evaluation methods,AEB system intervention time,relative distance,and braking deceleration were selected as evaluation indicators.The process of AEB comprehensive evaluation method was described,and an AEB system evaluation model was established based on the AEB test scenario set using analytic hierarchy process.The importance of elements was compared based on CIDAS accident casualty data,a judgment matrix was determined,evaluation indicators and scenario weight values were calculated,and consistency testing was performed.Combining the fuzzy evaluation method and applying the principle of maximum membership degree,the AEB system was comprehensively evaluated.(3)The hierarchical analytical fuzzy evaluation method was validated through virtual simulation testing.The advantages and disadvantages of the AEB algorithm were analyzed,and an AEB control strategy based on collision time was selected.A typical intersection and standard regulatory testing scenarios were constructed by using Prescan,and a collision-timebased AEB model was built by using Matlab&Simulink.Virtual simulation tests were conducted on different collision scenarios,such as car to car,car to motorcycle and car-topedestrian.Based on the performance of the AEB system in the testing scenarios,the AEB comprehensive evaluation method proposed in this paper was used to evaluate and validate the effectiveness and rationality of the method.Furthermore,the feasibility of evaluating autonomous emergency braking was demonstrated.