成对比较
场景测试
计算机科学
考试(生物学)
障碍物
任务(项目管理)
采样(信号处理)
避碰
测试用例
模拟
碰撞
工程类
人工智能
系统工程
机器学习
计算机安全
计算机视觉
回归分析
法学
古生物学
多样性(控制论)
滤波器(信号处理)
生物
政治学
作者
Yingying He,Yanfeng Wu,Jianping Gao
标识
DOI:10.1109/cac57257.2022.10055868
摘要
Aiming at the construction of driverless vehicle test scenario, this paper proposes a construction research of driverless vehicle test scenario library in campus environment. By analyzing the unique characteristics of the campus environment, the definition of different from the open/closed road test scenario is proposed. Combining with the characteristics of the scenario, the scenario elements are analyzed in layers. The combination analysis of scenario elements at different levels was carried out by using PICT(Pairwise Independent Combinatorial Testing) combination test tool, and random sampling method is used to extract different levels of sample sets. Through the analysis of scene elements and driving task complexity, this paper proposes a generation system of campus driverless vehicle combination test scenarios. According to the construction method of campus test scenario proposed, an obstacle avoidance test case is generated, and its simulation verification is carried out, the results show that the test vehicle can effectively decelerate and stop and avoid collision when encountering obstacles in front, which verifies the effectiveness of the test scenario.
科研通智能强力驱动
Strongly Powered by AbleSci AI