Adaptive Stress Testing with Reward Augmentation for Autonomous Vehicle Validatio
Adaptive Stress Testing with Reward Augmentation for Autonomous Vehicle Validatio
复制标题
DOI：
10.1109/itsc.2019.8917242
复制
发表时间：
2019-08
期刊：
影响因子：
--
通讯作者：
Anthony Corso;Peter Du;K. Driggs-Campbell;Mykel J. Kochenderfer
中科院分区：
文献类型：
--
作者：
Anthony Corso;Peter Du;K. Driggs-Campbell;Mykel J. Kochenderfer
确定可能的故障场景是评估自动驾驶汽车系统的关键步骤。真实的世界车辆测试通常用于自动驾驶车辆验证，但成本和时间要求很高。因此，模拟驱动的方法，如自适应压力测试（AST）已被提出来帮助验证。AST将寻找最可能的故障场景的问题表述为马尔可夫决策过程，可以使用强化学习来解决。在实践中，AST倾向于发现故障不可避免的场景，并且倾向于重复发现系统的相同类型的故障。这项工作解决了这些问题，通过编码域相关的信息到搜索过程中。通过这种修改，AST方法发现了一个更大的和更有表现力的子集的故障空间相比，原来的AST制定。我们表明，我们的方法是能够识别有用的自动驾驶汽车政策的失败场景。
Determining possible failure scenarios is a critical step in the evaluation of autonomous vehicle systems. Real world vehicle testing is commonly employed for autonomous vehicle validation, but the costs and time requirements are high. Consequently, simulation driven methods such as Adaptive Stress Testing (AST) have been proposed to aid in validation. AST formulates the problem of finding the most likely failure scenarios as a Markov decision process, which can be solved using reinforcement learning. In practice, AST tends to find scenarios where failure is unavoidable and tends to repeatedly discover the same types of failures of a system. This work addresses these issues by encoding domain relevant information into the search procedure. With this modification, the AST method discovers a larger and more expressive subset of the failure space when compared to the original AST formulation. We show that our approach is able to identify useful failure scenarios of an autonomous vehicle policy.