Automated scenario generation for regression testing of autonomous vehicles
Automated scenario generation for regression testing of autonomous vehicles
复制标题
用于自动驾驶车辆回归测试的自动场景生成
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
J. Seewig
中科院分区：
文献类型：
--
作者：
Elias Rocklage;Heiko Kraft;Abdullah Karatas;J. Seewig
自动驾驶汽车在技术上是可行的，正在成为现实。然而，在发射之前，必须进行彻底的测试。在本文中，我们提出了一种新的方法来自动生成测试场景的回归测试的自动驾驶汽车系统作为一个黑盒子在虚拟仿真环境中。为了实现这一点，我们专注于产生其他交通参与者的运动不失一般性的问题。我们结合联合收割机的组合交互测试方法与一个简单的轨迹规划器作为可行性检查，以产生有效的测试集与可变覆盖率。底层的约束满足问题，解决了一个简单的回溯算法。
Autonomous vehicles are technologically feasible and are becoming a reality. However, before they can be launched, thorough testing is necessary. In this paper, we present a novel approach to automatically generate test scenarios for regression testing of autonomous vehicle systems as a black box in a virtual simulation environment. To achieve this we focus on the problem of generating the motion of other traffic participants without loss of generality. We combine the combinatorial interaction testing approach with a simple trajectory planner as a feasibility checker to generate efficient test sets with variable coverage. The underlying constraint satisfaction problem is solved with a simple backtracking algorithm.
影响因子：
2.7
作者：
Geyer, Sebastian;Baltzer, Marcel;Winner, Hermann
通讯作者：
Winner, Hermann