基于小样本元学习算法的弹药装备质量预测

Small sample data-based meta-learning algorithm for equipment quality prediction

摘要

为了解决小样本数据下弹药装备质量预测困难等问题，该文提出了一种基于小样本元学习算法的弹药装备质量预测模型。首先，为了满足元学习模型的输入要求，将预处理的弹药装备质量数据分成多批次的任务序列；其次，通过一系列的学习任务来训练弹药装备质量元学习模型，以获得弹药装备质量预测知识，当处理从未遇到过的任务时，学习到的知识可以帮助提升新任务学习的适应性和概括性。最后，通过公开的实验数据集和弹药装备质量数据集来验证所提出的弹药装备小样本元学习模型在质量预测方面的优越性。

Abstract

To solve the difficulty of munition equipment quality prediction under small sample data, a munition equipment quality prediction model based on meta-learning algorithm is proposed in this paper. Firstly, the pre-processed ammunition equipment quality data is divided into multiple batches of task sequences to meet the input requirements of the small sample data meta-learning model. Secondly, the munition equipment quality meta-learning model is trained by a series of learning tasks to acquire the knowledge for predicting munition equipment quality, and when dealing with the tasks that have never been encountered before, the knowledge learned can help improve adaptability and generalization in learning these new tasks. Finally, the quality prediction validity of munition equipment quality meta-learning model is verified by open data set and munition equipment quality data set.

关键词

弹药装备质量预测 / 小样本数据 / 多批次任务序列 / 元学习

Key words

ammunition equipment quality / small sample data / multiple batches of task sequences / meta-learning

参考文献

[1] 赵乾, 郑娜娜, 李润伟, 等. 新形势下弹药产品质量检验的方法探析[J].新技术新工艺, 2022(5): 51-53.

[2] Zhao Qian, Zheng Nana, Li Runwei, et al. Analysis on methods of quality inspection of ammunition products under the new situation[J].New Technology & New Process, 2022(5): 51-53. (in Chinese)

[3] 李聪, 石义官, 黄权, 等. 基于深度学习的小口径弹药装配设备故障诊断专家系统[J].兵工自动化, 2023, 42(6): 20-26.

[4] Li Cong, Shi Yiguan, Huang Quan, et al. Fault diagnosis expert system of small caliber ammunition assembly equipment based on deep learning[J].Ordnance Industry Automation, 2023, 42(6): 20-26. (in Chinese)

[5] 康传国 . 面向对象的装备数据库系统设计及实现[D]. 成都: 电子科技大学,2014.

[6] Kang Chuanguo . Object-oriented weaponry Database system[D]. Chengdu: University of Electronic Science and Technology of China,2014. (in Chinese)

[7] 黄景德, 王军波, 王伟 . 小子样条件下制导弹药可靠性评估系统研究[J].弹箭与制导学报, 2006, 26(1): 535-536,541.

[8] Huang Jingde, Wang Junbo, Wang Wei . A study of guided ammunition reliability judging system based on little lots[J].Journal of Projectiles, Rockets, Missiles and Guidance, 2006, 26(1): 535-536,541. (in Chinese)

[9] Zhang Hongguang, Zhang Jing, Koniusz P. Few-shot learning via saliency-guided hallucination of samples[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach: IEEE, 2019: 2765-2774.

[10] Yao Huaxiu, Zhang Chuxu, Wei Ying, et al. Graph few-shot learning via knowledge transfer[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence. New York: AAAI Press, 2020: 6656-6663.

[11] Vinyals O, Blundell C, Lillicrap T, et al. Matching networks for one shot learning[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona: Curran Associates Inc., 2016: 3637-3645.

[12] Martin-Diaz I, Morinigo-Sotelo D, Duque-Perez O, et al. Early fault detection in induction motors using AdaBoost with imbalanced small data and optimized sampling[J].IEEE Transactions on Industry Applications, 2017, 53(3): 3066-3075.

[13] Mathew J, Pang C K, Luo Ming, et al. Classification of imbalanced data by oversampling in kernel space of support vector machines[J].IEEE Transactions on Neural Networks and Learning Systems, 2018, 29(9): 4065-4076.

[14] Saufi S R, Ahmad Z A B, Leong M S, et al. Gearbox fault diagnosis using a deep learning model with limited data sample[J].IEEE Transactions on Industrial Informatics, 2020, 16(10): 6263-6271.

[15] Peng Peng, Zhang Wenjia, Zhang Yi, et al. Cost sensitive active learning using bidirectional gated recurrent neural networks for imbalanced fault diagnosis[J].Neurocomputing, 2020, 407: 232-245.

[16] 毛万登, 袁少光, 姜亮, 等. 元学习优化的多重特征增强输变电缺陷检测网络[J].制造业自动化, 2025, 47(2): 59-67.

[17] Mao Wandeng, Yuan Shaoguang, Jiang Liang, et al. Multi-feature enhancement network for power transmission and transformation defect detection based on meta-learning optimization[J].Manufacturing Automation, 2025, 47(2): 59-67. (in Chinese)

[18] 黄毅, 郭金梦, 任子誉, 等. 基于改进元学习的小样本钢丝绳表面损伤检测[J].计算机工程与应用, 2025, 61(13): 256-269.

[19] Huang Yi, Guo Jinmeng, Ren Ziyu, et al. Few-shot surface damage detection of wire rope using improved meta-learning[J].Computer Engineering and Applications, 2025, 61(13): 256-269. (in Chinese)

[20] 李凌波, 田彦, 江旭东, 等. 一种基于元学习的改进YOLO钢管表面缺陷小样本检测模型[J].机电工程, 2025, 42(5): 985-993.

[21] Li Lingbo, Tian Yan, Jiang Xudong, et al. Improved meta-learning YOLO-based few-shot object detecting model for surface defects in steel tub[J].Journal of Mechanical & Electrical Engineering, 2025, 42(5): 985-993. (in Chinese)

[22] 宋阿隆, 崔学荣 . 基于元学习和轻量化注意力机制的小样本图像检索方法[J].物联网技术, 2025, 15(5): 71-74.

[23] Song A’long, Cui Xuerong . Small sample image retrieval method based on meta-learning and lightweight attention mechanism[J].Internet of Things Technologies, 2025, 15(5): 71-74. (in Chinese)

[24] 仲林林, 吴奇, 叶俊杰, 等. 基于元学习的变电设备小样本缺陷图像检测[J].仪器仪表学报, 2024, 45(10): 154-167.

[25] Zhong Linlin, Wu Qi, Ye Junjie, et al. Meta-learning-based few-shot image detection of defects in substation equipment[J].Chinese Journal of Scientific Instrument, 2024, 45(10): 154-167. (in Chinese)

[26] Finn C, Abbeel P, Levine S. Model-agnostic meta-learning for fast adaptation of deep networks[C]//Proceedings of the 34th International Conference on Machine Learning. Sydney: PMLR, 2017: 1126-1135.

[27] 郭佳, 郑山红, 陈闯, 等. 融合元学习和注意力机制的跨域推荐算法研究[J].计算机仿真, 2024, 41(12): 344-348.

[28] Guo Jia, Zheng Shanhong, Chen Chuang, et al. Research on cross-domain recommendation algorithm integrating meta-learning and attention mechanism[J].Computer Simulation, 2024, 41(12): 344-348. (in Chinese)

[29] Xu Zhixiong, Chen Xiliang, Cao Lei. Fast task adaptation based on the combination of model-based and gradient-based meta learning[J].IEEE Transactions on Cybernetics, 2022, 52(6): 5209-5218.

[30] Ren Chao, Jiang Bin, Lu Ningyun, et al. Meta-learning with distributional similarity preference for few-shot fault diagnosis under varying working conditions[J].IEEE Transactions on Cybernetics, 2024, 54(5): 2746-2756.

[31] 岳灵, 李晓宁, 韩楠, 等. 跨域图像空间数据少样本学习研究[J].无线电工程, 2024, 54(12): 2800-2819.

[32] Yue Ling, Li Xiaoning, Han Nan, et al. Research on cross-domain image spatial data few-shot learning[J].Radio Engineering, 2024, 54(12): 2800-2819. (in Chinese)

[33] Chen Jianjun, Hu Weihao, Cao Di, et al. A meta-learning method for electric machine bearing fault diagnosis under varying working conditions with limited data[J].IEEE Transactions on Industrial Informatics, 2023, 19(3): 2552-2564.

[34] Jose S T, Simeone O, Durisi G. Transfer meta-learning: Information-theoretic bounds and information meta-risk minimization[J].IEEE Transactions on Information Theory, 2022, 68(1): 474-501.

[35] 简丽琼 . 多模态智能中的跨模态迁移学习策略研究[J].信息与电脑, 2024, 36(22): 180-182.

[36] Jian Liqiong . Research on cross-modal transfer learning strategy in multi-modal intelligence[J].Information & Computer, 2024, 36(22): 180-182. (in Chinese)

[37] Tian Pinzhuo, Li Wenbin, Gao Yang. Consistent meta-regularization for better meta-knowledge in few-shot learning[J].IEEE Transactions on Neural Networks and Learning Systems, 2022, 33(12): 7277-7288.

[38] 牛卉, 贾寓词, 李浩成, 等. 舰载机用航空弹药装备情况研究[J].国防科技, 2022, 43(5): 29-34.

[39] Niu Hui, Jia Yuci, Li Haocheng, et al. The equipment of aerial bombs for shipborne aircraft[J].National Defense Technology, 2022, 43(5): 29-34. (in Chinese)

[40] Kolarik M, Burget R, Riha K. Comparing normalization methods for limited batch size segmentation neural networks[C]//Proceedings of the 43rd International Conference on Telecommunications and Signal Processing. Milan: IEEE, 2020: 677-680.

[41] 张江鹏, 付云伟, 吴纯治, 等. 面向多工况设备故障诊断的改进灰色关联分析法[J].机械, 2023, 50(6): 67-73.

[42] Zhang Jiangpeng, Fu Yunwei, Wu Chunzhi, et al. Improved grey correlation analysis method for fault diagnosis of multi-mode equipment[J].Machinery, 2023, 50(6): 67-73. (in Chinese)

[43] Zhang Peng, Liu Chenbing, Chang Xingshuo, et al. Metric-based meta-learning model for few-shot PolSAR image terrain classification[C]//Proceedings of the CIE International Conference on Radar. Haikou: IEEE, 2021: 2529-2533.

[44] 胡磊, 赵辉, 南熠, 等. 基于搜索规则和交叉熵优化的无人机路径规划方法[J].电子与信息学报, 2023, 45(6): 2144-2152.

[45] Hu Lei, Zhao Hui, Nan Yi, et al. Unmanned aerial vehicle path planning method based on search rule and cross entropy optimization[J].Journal of Electronics & Information Technology, 2023, 45(6): 2144-2152. (in Chinese)

[46] 王卓鑫, 赵海涛, 谢月涵, 等. 反向传播神经网络联合遗传算法对复合材料模量的预测[J].上海交通大学学报, 2022, 56(10): 1341-1348.

[47] Wang Zhuoxin, Zhao Haitao, Xie Yuehan, et al. Prediction of modulus of composite materials by BP neural network optimized by genetic algorithm[J].Journal of Shanghai Jiao Tong University, 2022, 56(10): 1341-1348. (in Chinese)

[48] 刘美玉, 刘启发, 程亚玲, 等. 基于梯度下降算法的光学神经网络性能的研究[J].微电子学与计算机, 2022, 39(12): 13-20.

[49] Liu Meiyu, Liu Qifa, Cheng Yaling, et al. Research on the performance of optical neural network based on gradient descent algorithm[J].Microelectronics & Computer, 2022, 39(12): 13-20. (in Chinese)

[50] Minoofam S A H, Bastanfard A, Keyvanpour M R. TRCLA: A transfer learning approach to reduce negative transfer for cellular learning automata[J].IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(5): 2480-2489.

[51] Fritz M, Hayman E, Caputo B, et al. The kth-tips database[EB/OL]. Stockholm: KTH Royal Institute of Technology,2004[2025-03-20].https://www.csc.kth.se/cvap/databases/kth-tips/index.html.

基金资助

国防基础科研项目(JCKY2022209A002)