为了解决基于深度学习的列车智能视觉定位系统难以测试问题，提出一种面向列车智能视觉定位的安全测试方法。基于风格迁移思想，通过构建生成式对抗网络（GAN）实现测试用例的生成；基于深度变异测试方法，实现对测试用例错误检测能力的量化评价；针对城轨运营组织特点，提出一种“虚拟-半实-真实”平行测试平台架构，用于支持测试用例生成模型的构建和测试执行。实验结果表明，本方法生成的测试用例场景种类分布更为均匀多样，能够较为全面地测试模型在不同场景下的安全性，有效提升列车智能视觉定位的测试效率。
In order to solve the problem that intelligent train visual positioning system based on deep learning is difficult to test, this paper proposes a safety test method for intelligent train visual positioning. Firstly, based on the idea of Image-to-Image translation, we construct a generative adversarial network (GAN) to generate test cases. Then we implement the quantitative evaluation of the error detection ability of test cases based on deep mutation testing. Finally, according to the characteristics of urban rail operation organization, we propose a parallel test platform architecture of "virtual-reality, semi-reality, reality" to support the construction of the test case generation model and test execution. The method proposed in this paper provides a basis for ensuring the safety of intelligent visual train positioning, provides a new research idea for the safety application of intelligent visual perception technology in the autonomous running of trains, and plays an essential role in ensuring the safety of trains.
[1] 曹启滨 . 城市轨道交通列车定位方法分析[J]. 铁路通信信号工程技术, 2012, 9(1): 55-56, 62.
[2] 李杰 . 基于组合定位的下一代列车自主定位系统研究[J]. 城市轨道交通研究, 2020, 23(11): 93-95.
[3] 杨岗, 林颖 . 基于多传感器的信息融合列车定位算法研究[J]. 铁道通信信号, 2019, 55(5): 42-47.
[4] 王小可, 孔青宁. 基于5G同步信号的高速列车定位方法[J]. 计算机测量与控制, 2021, 29(5): 159-163.
[5] 万赟 . 从图灵测试到深度学习: 人工智能 60 年[J]. 科技导报, 2016, 34(7): 26-33.
[6] Wang Z Y, Yu G Z, Zhou B, et al. A train positioning method based-on vision and millimeter-wave radar data fusion[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(5): 4603-4613.
[7] 滕达, 赵阳, 范楷, 等. 列车自主定位图像识别技术应用研究[J]. 铁道运输与经济, 2020, 42(12): 43-48.
[8] Yan H R, Zhao P P, Zhuang F Z, et al. Cross-domain recommendation with adversarial examples[M]//Database Systems for Advanced Applications. Cham: Springer International Publishing, 2020: 573-589.
[9] Kang D, Raghavan D, Bailis P, et al. Model assertions for monitoring and improving ML models[DB/OL]. arXiv preprint: 2003.01668, 2022.
[10] Yuan Q, Peng Y, Xu X D, et al. Key points of investigation and analysis on traffic accidents involving intelligent vehicles[J]. Transportation Safety and Environment, 2021, 3(4): tdab020.
[11] Jenssen G D, Moen T, Johnsen S O. Accidents with Automated Vehicles-Do self-driving cars need a better sense of self[C]//Proceedings of the 26th ITS World Congress. Singapore: ITS. 2019: 21-25.
[12] Cho H S. Operational design domain (ODD) framework for driver-automation integrated systems[D]. Cambridge: Massachusetts Institute of Technology, 2020.
