Virtual Test Scenarios Development For Highway Straight Driving Condition

Scenario-based virtual testing of autonomous driving has great advantages in test cost and efficiency,and has become a research hotspot in scientific research institutions and universities.Highway operating conditions are an important part of autonomous driving scenarios.At present,domestic and foreign The research is still in its infancy,so the research is of great significance.This article focuses on the core issues of highway automatic driving test scenarios-test scenarios,accelerated test methods and automated virtual test platforms.According to the definition and element analysis of the test scenarios,the test scenarios are classified and designed separately,and reasonable accelerated test methods are selected for comparison.,To build an automated generation platform for virtual test scenes,and finally realize the automatic generation of virtual test scenes for highway automatic driving.First,by combing through a large number of references,the test scene is divided into a static scene part and a dynamic scene part,a reasonable acceleration test method is selected for the generation of the static scene,and the static scene and the dynamic scene are designed separately.The static scene part includes the weather environment and the static road network structure.In order to realize the generation of static scene test cases,the relevant factors affecting the automatic driving performance in the static scene are analyzed and parameterized,and the influencing factors are divided into discrete factors(number of lanes,Guardrail types,etc.)and continuous factors(time,visibility,etc.).According to the type of highway road structure,the discrete influencing factors are combined with each other to form a static road scene set,which is built in the VTD(Virtual Test Drive)software.Discretize continuous influencing factors,and add weights and constraints to them according to actual road conditions.In the accelerated test method,the advantages and disadvantages of the test case generation method are compared,a combination test method with high repeatability and strong coverage is selected,and the static road scene set and continuous factors are used as the combination test input parameters to calculate,and the static scene test is finally obtained Set of use cases.The dynamic scene part is an environmental vehicle,and its micro-driving characteristics and macro-generating characteristics are studied separately.In terms of micro-driving characteristics,in order to meet the driving characteristics of real highway vehicles,first classify real highway driving car-following conditions,design virtual test scenarios,build a driving simulation platform,and collect test driver car-following data.Then pre-process the collected data and establish a simplified car following model based on the actual car following process.At the same time,it classifies the driver’s driving style based on the collected data.Finally,multiple linear fitting is performed on the car-following process of the same type of drivers,and the driving mode judgment method is set to obtain a micro driving model based on the real driver’s style.In terms of the macroscopic generation characteristics,a macroscopic traffic flow headway model is fitted based on real road traffic data sets and used as a vehicle macroscopic generation model.Finally,build a micro driving model in Simulink.Based on VTD,Simulink and NI-PXI,build an automated generation platform for virtual test scenarios,write corresponding data interaction interfaces,deploy static and dynamic scenarios to the automated generation platform,and complete the automated generation of virtual test scenarios for highway straight-line driving conditions.