基于反馈特征提取和因果反演辨识的自主智能系统效能评估方法

doi:10.12305/j.issn.1001-506X.2026.01.19

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (1): 209-217.doi: 10.12305/j.issn.1001-506X.2026.01.19

基于反馈特征提取和因果反演辨识的自主智能系统效能评估方法

余晨浩¹^,²(), 常雷雷¹^,², 周宇²^,*, 孙建彬³

1. 杭州电子科技大学自动化学院，浙江杭州 310018
2. 西安电子科技大学协同智能系统教育部重点实验室，陕西西安 710126
3. 国防科技大学系统工程学院，湖南长沙 410073

收稿日期:2024-10-08 出版日期:2026-01-25 发布日期:2026-02-11
通讯作者: 周宇 E-mail:231060041@hdu.edu.cn
作者简介:余晨浩（2001—）男，博士研究生，主要研究方向为数据挖掘、机器学习、复杂系统设计
常雷雷（1985—），男，副研究员，博士，主要研究方向为复杂作战条件下的决策分析及能力评估
孙建彬（1989—），男，副教授，博士，主要研究方向为体系能力评估、体系发展规划
基金资助:
国家自然科学基金（72471067，72471238）资助课题

Performance evaluation method for autonomous intelligent system using feedback-based attribute extraction and anti-causality

Chenhao YU¹^,²(), Leilei CHANG¹^,², Yu ZHOU²^,*, Jianbin SUN³

1. School of Automation，Hangzhou Dianzi University， Hangzhou 310018， China
2. Key Laboratory of Collaborative Intelligence Systems, Ministry of Education，Xidian University， Xi’an 710126 China
3. College of Systems Engineering，National University of Defense Technology， Changsha 410073， China

Received:2024-10-08 Online:2026-01-25 Published:2026-02-11
Contact: Yu ZHOU E-mail:231060041@hdu.edu.cn

摘要/Abstract

摘要：

为解决以无人系统为代表的自主智能系统在实际和仿真实验中产生的海量数据特征超载和因果关系不一致的问题，提出基于反馈闭环策略的特征提取方法。首先，在传统特征约简基础上，反向提取原始数据空间中的特征。然后，基于因果反演的数据辨识方法，通过计算数据因果逻辑关系一致性来辨识并排除与主要数据因果关系不一致的因果反演数据。最后，以无人系统作为示例的验证结果表明，在仅保留54.14%原始信息的情况下，系统效能评估误差降低约12.43%。所提方法具有较强的泛化能力，以及模型独立性和鲁棒性。

关键词: 反馈特征提取, 因果反演辨识, 自主智能系统, 效能评估

Abstract:

To solve the problem of massive data feature overload and anti-causal data in real and simulated tests, a feature extraction method is proposed based on feedback closed-loop strategy. Firstly, extracts features from the original data space in reverse based on traditional feature reduction. Secondly, data identification method based on anti-causality identification, which identifies and excludes anti-causality data by calculating the consistency of causal logical relationships of data. Finally, the verification results using unmanned systems as an example show that, while retaining only 54.14% of the original information, system performance evaluation error reduced by approximately 12.43%. The method proposed has strong generalization ability, model independence and robustness.

Key words: feedback-based attribute extraction, anti-causality identification, autonomous intelligent system, performance evaluation

中图分类号:

TP 18

余晨浩, 常雷雷, 周宇, 孙建彬. 基于反馈特征提取和因果反演辨识的自主智能系统效能评估方法[J]. 系统工程与电子技术, 2026, 48(1): 209-217.

Chenhao YU, Leilei CHANG, Yu ZHOU, Jianbin SUN. Performance evaluation method for autonomous intelligent system using feedback-based attribute extraction and anti-causality[J]. Systems Engineering and Electronics, 2026, 48(1): 209-217.

图/表 9

图1

图2

图3

图4

图5

表1

表2

图6

表3

参考文献 30

1	CHEN J, SUN J, WANG G. From unmanned systems to autonomous intelligent systems[J]. Engineering, 2022, 12, 16- 19. doi: 10.1016/j.eng.2021.10.007
2	TAN Y W, WANG J D, LIU J J, et al. Unmanned systems security: models, challenges, and future directions[J]. IEEE Network, 2020, 34 (4): 291- 297. doi: 10.1109/MNET.001.1900546
3	MOHSAN S A H, KHAN M A, NOOR F, et al. Towards the unmanned aerial vehicles （UAVs）: a comprehensive review[J]. Drones, 2022, 6 (6): 147. doi: 10.3390/drones6060147
4	WANG H F, CHENG H J, HAO H Y. The use of unmanned aerial vehicle in military operations[C]//Proc. of the International Conference on Man-Machine-Environment System Engineerin, 2020.
5	HYLAND S L, FALTYS M, HUSER M, et al. Early prediction of circulatory failure in the intensive care unit using machine learning[J]. Nature Medicine, 2020, 26, 364- 373. doi: 10.1038/s41591-020-0789-4
6	BAKKER B, ZABłOCKI B, BAKER A, et al. A multi-stage, multi-feature machine learning approach to detect driver sleepiness in naturalistic road driving conditions[J]. IEEE Trans. on Intelligent Transportation Systems, 2021, 23 (5): 4791- 4800.
7	FU S X, LI H X, LIU Y, et al. Social media overload, exhaustion, and use discontinuance: examining the effects of information overload, system feature overload, and social overload[J]. Information Processing & Management, 2020, 57 (6): 102307.
8	YANG M J, LIM S Y, PARK H J, et al. Solving the data overload: device-to-device bearer control architecture for cellular data offloading[J]. IEEE Vehicular Technology Magazine, 2013, 8 (1): 31- 39. doi: 10.1109/MVT.2012.2234052
9	PAWLOWSKI N, CASTRO D C, GLOCKER B. Deep structural causal models for tractable counterfactual inference[J]. Advances in Neural Information Processing Systems, 2020, 33, 857- 869.
10	ANOGUEIR A R, GAMA J, FERREIRA C A. Causal discovery in machine learning: theories and applications[J]. Journal of Dynamics & Games, 2021, 8 (3): 203- 231.
11	PROSPERI M, GUO Y, SPERRIN M, et al. Causal inference and counterfactual prediction in machine learning for actionable healthcare[J]. Nature Machine Intelligence, 2020, 2, 369- 375. doi: 10.1038/s42256-020-0197-y
12	POPESCU D C, DUMITRACHE I. Knowledge representation and reasoning using interconnected uncertain rules for describing workflows in complex systems[J]. Information Fusion, 2023, 93, 412- 428. doi: 10.1016/j.inffus.2023.01.007
13	WANG J R, LIU Y, LI P G, et al. Overview of data quality: examining the dimensions, antecedents, and impacts of data quality[J]. Journal of the Knowledge Economy, 2024, 15, 1159- 1178. doi: 10.1007/s13132-022-01096-6
14	HERNAN M A, WANG W, LEAF D E. Target trial emulation: a framework for causal inference from observational data[J]. JAMA, 2022, 328 (24): 2446- 2447. doi: 10.1001/jama.2022.21383
15	CHANG L L, ZHOU Z J, YOU Y, et al. Belief rule based expert system for classification problems with new rule activation and weight calculation procedures[J]. Information Sciences, 2016, 336, 75- 91. doi: 10.1016/j.ins.2015.12.009
16	HOU Z Q, MARTINEZ-GARCIA M, ZHANG Y, et al. Entropy measures in machine fault diagnosis: Insights and applications[J]. IEEE Trans. on Instrumentation and Measurement, 2020, 69 (6): 2607- 2620. doi: 10.1109/TIM.2020.2981220
17	SANG B B, CHEN H M, YANG L, et al. Incremental feature selection using a conditional entropy based on fuzzy dominance neighborhood rough sets[J]. IEEE Trans. on Fuzzy Systems, 2021, 30 (6): 1683- 1697.
18	BEIRANVAND F, MEHRDAD V, DOWLATSHAHI M B. Unsupervised feature selection for image classification: a bipartite matching-based principal component analysis approach[J]. Knowledge-Based Systems, 2022, 250, 109085. doi: 10.1016/j.knosys.2022.109085
19	ANOWAR F, SADAOUI S, SELIM B. Conceptual and empirical comparison of dimensionality reduction algorithms （PCA, KPCA, LDA, MDS, SVD, LLE, ISOMAP, LE, ICA, t-SNE）[J]. Computer Science Review, 2021, 40, 100378. doi: 10.1016/j.cosrev.2021.100378
20	LIANG X P, TANG Z J, WU J L, et al. Robust image hashing with isomap and saliency map for copy detection[J]. IEEE Trans. on Multimedia, 2021, 25, 1085- 1097.
21	WU X, HONG D F, CHANUSSOT J. Convolutional neural networks for multimodal remote sensing data classification[J]. IEEE Trans. on Geoscience and Remote Sensing, 2021, 60
22	张效康. 地理国情监测数据可靠性分析与控制方法研究[D]. 武汉: 武汉大学, 2017.
	ZHANG X K. Reliability analysis and controlling methods for national geographic state monitoring data [D]. Wuhan: Wuhan University, 2017.
23	YU C H, CHANG L L, XU X B, et al. Compatible multi-model output fusion for complex systems modeling via set operation-based focus data identification[J]. Journal of Computational Science, 2024, 82, 102423. doi: 10.1016/j.jocs.2024.102423
24	YU C H, ZHU J Y, CHANG L L, et al. A novel data credibility-centric multi-model-based complex systems modeling approach for UAV capability evaluation[J]. International Journal of Machine Learning and Cybernetics, 2025, 16, 2687- 2704. doi: 10.1007/s13042-024-02415-w
25	JAVED S, HASSAN A, AHMAD R, et al. State-of-the-art and future research challenges in uav swarms[J]. IEEE Internet of Things Journal, 2024, 11 (11): 19023- 19045. doi: 10.1109/JIOT.2024.3364230
26	陈士涛, 张海林. 基于作战网络模型的异构无人机集群作战能力评估[J]. 军事运筹与系统工程, 2019, 33 (1): 38- 43.
	CHEN S T, ZHANG H L. Evaluation of heterogeneous UAV cluster combat capability based on combat network model[J]. Military Operations Research and Systems Engineering, 2019, 33 (1): 38- 43.
27	赵蕊蕊, 于海跃, 游雅倩, 等. 无人集群试验评估现状及技术方法综述[J]. 系统工程与电子技术, 2024, 46 (2): 570- 585.
	ZHAO R R, YU H Y, YOU Y Q, et al. Review on current status and technology method of unmanned swarm test evaluation[J]. Systems Engineering and Electronics, 2024, 46 (2): 570- 585.
28	范波, 钟季龙, 徐丽霞, 等. 基于因果熵的无人集群对抗评估指标分配方法[J]. 系统工程与电子技术, 2024, 46 (6): 2034- 2043.
	FAN B, ZHONG J L, XU L X, et al. Index allocation method for unmanned swarm confrontation evaluation based on causal entropy[J]. Systems Engineering and Electronics, 2024, 46 (6): 2034- 2043.
29	SUN C R, FONTANESI G, CANBERK B, et al. Advancing UAV communications: a comprehensive survey of cutting-edge machine learning techniques[J]. IEEE Open Journal of Vehicular Technology, 2024, 5, 825- 854. doi: 10.1109/OJVT.2024.3401024
30	BAI L, LIANG J Y. K-relations-based consensus clustering with entropy-norm regularizers[J]. IEEE Trans. on Neural Networks and Learning Systems, 2024, 35 (12): 17662- 17673. doi: 10.1109/TNNLS.2023.3307158

模型	模型输入（训练集）			测试集结果MAE
模型	特征数量	数据量	输入信息百分比/%	测试集结果MAE
基准模型	16 （100%）	400 （100%）	100	范围	[0.0750, 0.1972]
				均值	0.1360
				方差	1.2416E-03
特征提取后模型	9 （56.25%）	400 （100%）	56.25	范围	[0.0864, 0.2058]
				均值	0.1388
				方差	1.3285E-03
数据辨识后模型	16 （100%）	385 （96.25%）	96.25	范围	[0.0541, 0.1669]
				均值	0.1110
				方差	7.9503E-04
特征提取+数据辨识后模型	9 （56.25%）	385 （96.25%）	54.14	范围	[0.0831, 0.1741]
				均值	0.1191
				方差	4.9817E-04

不同方法	训练集输入		测试集结果（MAE）
不同方法	特征数量	数据量	测试集结果（MAE）
基准模型	16	400	均值	0.1360
特征选择信息熵	9	400	均值	0.1510
特征选择Isomap	9	400	均值	0.1530
数据辨识（排除5组）	16	395	均值	0.1402
数据辨识（排除15组）	16	385	均值	0.1427
数据辨识（排除30组）	16	370	均值	0.1643
数据辨识（KRCC）	16	380	均值	0.1420

基准模型	模型	模型的输入（训练集）				测试集MAE
基准模型	模型	特征数量	数据量	百分比/%	均值	范围	方差	变化/%
BPNN	Baseline	16	400	100	0.1360	[0.0750,0.1972]	1.2416E-03	—
	Model_att	9	400	56.25	0.1388	[0.0864,0.2058]	1.3285E-03	2.06
	Model_quan	16	385	96.25	0.1110	[0.0541,0.1669]	7.9503E-04	−18.38
	Model_both	9	385	54.14	0.1191	[0.0831,0.1741]	4.9817E-04	−12.43
RBF	Baseline	16	400	100	0.1403	[0.0592,0.1991]	1.0032E-03	—
	Model_att	9	400	56.25	0.1441	[0.0976,0.1771]	4.3198E-04	2.71
	Model_quan	16	385	96.25	0.1171	[0.0687,0.1860]	8.3103E-04	−16.54
	Model_both	9	385	54.14	0.1227	[0.0611,0.1819]	5.9187E-04	−12.55
GPR	Baseline	16	400	100	0.1816	[0.1126,0.2346]	5.0565E-04	—
	Model_att	9	400	56.25	0.1858	[0.1460,0.2183]	7.0347E-04	2.31
	Model_quan	16	385	96.25	0.1603	[0.0975,0.2373]	5.4856E-04	−11.73
	Model_both	9	385	54.14	0.1655	[0.1040,0.1813]	3.9483E-04	−8.87

[1]	孙文, 程娟, 王一豪, 曹开臣, 葛萌萌, 黄刘, 张耿. 高光谱卫星舰船目标检测效能分析与建模研究[J]. 系统工程与电子技术, 2025, 47(5): 1432-1442.
[2]	陈顶, 方志耕, 杨保华, 叶丰, 张娜, 张靖如. 考虑指标协同效应重构的联合作战体系效能评估灰色主成分模型[J]. 系统工程与电子技术, 2025, 47(5): 1561-1574.
[3]	袁彰求, 杨朝旭, 荣海军. 面向定性与定量指标的轻量化高空飞艇效能评估方法[J]. 系统工程与电子技术, 2025, 47(3): 817-826.
[4]	韩维, 郭放, 刘玉杰, 苏析超, 刘洁. 基于三角模糊数作战环的航母编队作战效能评估方法[J]. 系统工程与电子技术, 2025, 47(3): 893-903.
[5]	孙志国, 刘传令, 王震铎. 基于混合双链量子遗传算法的干扰效能评估方法[J]. 系统工程与电子技术, 2025, 47(12): 4186-4195.
[6]	朱自强, 刘松涛, 彭锐晖. 基于Vague拓扑融合和改进博弈论的雷达侦察系统作战效能评估[J]. 系统工程与电子技术, 2025, 47(11): 3699-3707.
[7]	郑丽莎, 尹东亮, 王旋. 基于改进D-S证据理论的相控阵雷达作战效能评估[J]. 系统工程与电子技术, 2024, 46(4): 1330-1336.
[8]	马力, 师鹏, 陈宇, 李文龙. 天基信息支援体系离散事件仿真与效能评估[J]. 系统工程与电子技术, 2024, 46(3): 906-913.
[9]	胡涛, 申立群, 朱镜达, 孙成会, 董伟锋. 基于FAST和Sobol指数法的雷达系统效能敏感性分析[J]. 系统工程与电子技术, 2024, 46(2): 561-569.
[10]	陈宇, 师鹏, 马力, 李文龙. 天基信息支援体系建模与效能评估方法[J]. 系统工程与电子技术, 2024, 46(10): 3407-3415.
[11]	曹嘉平, 欧萌歆, 李易珊, 姜江, 李际超. 岛礁防空电子对抗装备体系构建与效能评估[J]. 系统工程与电子技术, 2023, 45(9): 2784-2792.
[12]	浣顺启, 方哲梅, 王剑波. 基于功能依赖网的体系效能评估方法[J]. 系统工程与电子技术, 2022, 44(7): 2191-2200.
[13]	蔺向阳, 邢清华, 刘付显. 针对要点防空模型的作战兵力优化研究[J]. 系统工程与电子技术, 2022, 44(3): 921-928.
[14]	邱禄芸, 方志耕, 陶良彦, 陶秋澄. 网络体系效能评估改进FDNA模型[J]. 系统工程与电子技术, 2022, 44(12): 3728-3737.
[15]	潘星, 张振宇, 张艳梅, 王冉冉. 基于Sobol敏感性分析的装备体系保障效能评估[J]. 系统工程与电子技术, 2021, 43(2): 390-398.

基于反馈特征提取和因果反演辨识的自主智能系统效能评估方法

Performance evaluation method for autonomous intelligent system using feedback-based attribute extraction and anti-causality

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 30

相关文章 15

编辑推荐

Metrics

本文评价