Research On Critical Scenarios Generation Method And Simulation Test Considering The Effect Of Longitudinal Slope

Scenario testing is a key technology in the field of autonomous driving.The research of test scenario generation methods is an important direction for the expected functional safety of autonomous driving.Because of the randomness of natural driving scenarios,critical and dangerous test scenarios become low-probability events in autonomous driving tests,resulting in the inability to fully verify the safety of autonomous vehicles.As a significant part of automatic driving,the automatic emergency braking system is an important part of automatic driving,and the relevant research on its test scenarios has become mature.Therefore,there is a higher demand for research on critical scenarios generation methods for automatic emergency braking under longitudinal slopes.Although the current method has achieved great results in scene analysis and rapid generation of critical scenarios,how to determine the reasonable domain of the scene based on scene analysis and how to combine it with the search method achieve faster finding of the target scene still needs further research.In this regard,this paper studies the critical scenarios generation method of automatic emergency braking system considering the influence of longitudinal slope,and verifies it in the simulation environment.The main research contents are as follows:The constraint relationship among the scene parameters of longitudinal slope emergency braking is analyzed.Based on the real ramp vehicle driving data,by comparing the differences in the characteristics of the transition of the driving state under different longitudinal slopes,the influence of the road slope on the driving behavior is demonstrated from the data;and then the emergency braking scene on the ramp is transformed into a scene consisting of 7parameters Based on the specific test cases of the vehicle driving dynamics equation,the dynamic constraints of the scene parameters are found;the parameters of the vehicle state before and after the scene are found based on the automatic emergency braking requirements.The constraint relationship of the scene parameters provides a prerequisite for obtaining the reasonable domain of the scene.Proposed a key use case generation method for automatic emergency braking on longitudinal slopes based on scene rational domain evolution.Combining with the idea of evolution,in the reasonable domain of the scene,the satisfaction criterion model can be combined with the genetic algorithm to achieve more efficient target search.Based on the satisfaction criterion model,this study designs a critical scenarios objective function with critical collision as the goal.Based on the scene parameters Constraint priority,design evolution operators that satisfy the constraint relationship of scene parameters,including crossover operator and mutation operator;from the aspects of the efficiency of critical scenarios generation and scene coverage,compared the current test scene generation based on traditional genetic algorithm method,and analyzed the influence of different road slopes on the danger zone of the scene.Based on the co-simulation of PreScan and matlab,the generated scene was verified.Based on the simulation environment,the key scenarios of automatic emergency braking on longitudinal slopes are analyzed,and the scene model of emergency braking on longitudinal slopes is found.According to the automatic emergency braking function test requirements,the slope automatic emergency braking control model is established,and based on the slope automatic emergency braking control model,the critical scenarios test cases of longitudinal slope automatic emergency braking are verified in the simulation environment.