Research On Test And Evaluation Method Of Autonomous Vehicle Emergency Braking System Based On Typical Test Scenarios

Autonomous vehicles are intelligent vehicles that integrate environmental perception,decision-making and planning,and control execution,and represent the future development direction of automotive technology.Automatic Emergency Braking System(AEB),as an active safety system that can actively detect collision risks and autonomously brake to avoid collisions,is an important part of autonomous vehicles below L3 and is currently the most widely used auxiliary driving system.The real traffic environment is complex and changeable.Testing and evaluating the safety and effectiveness of the AEB system is a necessary prerequisite for its large-scale application.However,the current domestic and foreign AEB test standards and regulations and test methods are mostly oriented to straight test scenarios,and the test results mostly adopt subjective evaluation methods based on expert experience,and lack AEB test methods suitable for complex road environments and AEB comprehensive evaluation methods under multi-scenario conditions.This paper analyzes the existing AEB test and evaluation methods,refers to the domestic and foreign AEB test and evaluation standards and regulations,and designs the AEB test scenarios and test methods in the crossroad environment according to the characteristics of urban roads,and forms the typical AEB test scenarios together with the legal straight road scenarios.AEB test method under typical test scenarios.On this basis,this paper proposes an AEB comprehensive evaluation method that combines subjectivity and objectiveness,constructs a simulation and real-vehicle test system,and realizes AEB simulation test evaluation and real-vehicle test analysis for typical test scenarios.The main research work of the thesis is as follows:Construct typical AEB test scenarios and test methods.Based on the analysis of the current domestic and foreign AEB test standards and regulations,starting from the actual situation of China’s urban road traffic,this paper designs the AEB test scenario in the crossroad environment with the vehicle turning left at the intersection as the core,covering the vehicle’s There are multiple collision risks that may be encountered when turning,and corresponding test methods are proposed in combination with the requirements of test regulations.A typical AEB test scenario is constructed using the intersection test scenario designed in this paper and the straight road test scenario set in the regulations.An AEB subjective and objective comprehensive evaluation method oriented to typical test scenarios is proposed.The evaluation method takes the relative distance of the target,the collision time TTC,and the braking deceleration as indicators,and establishes the AEB three-layer evaluation index model,and determines the weight of each evaluation index and each test scene through the analytic hierarchy process.Applying the principle of fuzzy change and the principle of maximum membership degree,the test evaluation in a single test scenario of the AEB system and the comprehensive quantitative evaluation in all typical test scenarios are realized,and the AEB evaluation process under the typical test scenario is improved.The AEB test system is constructed based on the Prescna simulation environment to realize AEB test evaluation in typical test scenarios.Use Prescan to build the AEB straight test scene and the left turn test scene at the intersection.Three kinds of AEB control algorithms are constructed using Matlab to realize the Prescan-Matlab joint simulation test.According to the evaluation method proposed in this paper,the simulation test results under typical AEB test scenarios are comprehensively evaluated.The simulation results show that: the designed typical test scenarios and test methods meet the AEB system test requirements;the proposed comprehensive evaluation method can make a comprehensive quantitative evaluation of the performance of the AEB system equipped with three different algorithms.The final evaluation result shows that it is based on TTC The final score of the AEB system of the algorithm is 78.97,the final score of the AEB system based on the Mazda algorithm is 64.63,and the final score of the AEB system based on the Honda algorithm is 56.65,so it can be seen that the TTC algorithm is better than the other two algorithms.Based on the closed test field,build the AEB real-vehicle test system and carry out real-vehicle tests.According to the actual vehicle field test requirements,the architecture of the actual vehicle field test system was designed,and the design schemes for dummy targets,pedestrian drive systems,differential positioning systems,and upper computer control systems were proposed.The actual vehicle test system was carried out based on the constructed AEB real vehicle test system.Car test.The test results show that the actual vehicle field test can be close to the real traffic road environment,and obtain the operating state of the vehicle in the actual traffic environment,which provides further verification for the theoretical method proposed in this paper.