基于Bayesian-SEM的特种车辆乘员决策行为影响因素研究

doi:10.16265/j.cnki.issn1003-3033.2026.07.0602

摘要/Abstract

摘要：

为探究特种车辆乘员火控任务决策行为的关键影响因素及作用机制,基于信息处理过程将任务解构为目标感知、态势判定、火力决控3个阶段,综合运用层次分析法(AHP)与贝叶斯结构方程模型(Bayesian-SEM)构建多因素影响路径模型。通过文献分析与任务解构,从内源性与外源性2个维度初步提取9项影响因素,采用AHP法结合5位专家评估筛选出时间压力、信息动态性、任务复杂度与心理应激4项关键变量。基于235份有效问卷建立Bayesian-SEM,采用马尔可夫链蒙特卡罗采样进行参数估计,并通过多群组检验验证模型跨群体稳定性。结果表明:信息动态变化对心理应激和时间压力均呈显著正向影响;任务复杂度与时间压力呈负向关联,形成“任务复杂度→时间压力→决策水平”的非典型作用路径;心理应激对决策水平呈正向影响,时间压力则负向作用于决策水平。与传统SEM相比,Bayesian-SEM在有限样本下参数估计标准误更小、稳定性更优。

关键词: 特种车辆乘员, 贝叶斯结构方程模型(Bayesian-SEM), 火控任务, 层次分析法(AHP), 决策行为影响因素

Abstract:

To investigate the key influencing factors and action mechanisms of decision-making behavior among special vehicle crew members during fire control tasks, the tasks were decomposed into three stages—target perception, situation assessment, and firepower decision-control—based on the information processing workflow. A multi-factor influence path model was constructed by integrating AHP and Bayesian-SEM. Through literature analysis and task decomposition, nine preliminary influencing factors were extracted from the endogenous and exogenous dimensions. Four key variables were screened using the AHP method combined with evaluations from five experts. These variables included time pressure, information dynamics, task complexity, and psychological stress. Based on 235 valid questionnaire responses, a Bayesian-SEM model was established. Parameters were estimated using Markov Chain Monte Carlo sampling. The cross-group stability of the model was verified through multi-group testing. The results show that dynamics changes in information have significant positive effects on both psychological stress and time pressure. A negative association is observed between task complexity and time pressure. Thus an atypical pathway of "task complexity → time pressure → decision-making level" is formed. Psychological stress has a positive effect on decision-making level. Time pressure has a negative effect on decision-making level. Compared with traditional SEM, Bayesian-SEM has smaller standard errors and better stability in parameter estimation under limited sample conditions.

Key words: occupants of special vehicles, Bayesian structural equation modeling (Bayesian-SEM), fire control mission, analytic hierarchy process (AHP), influencing factors of decision-making behavior

中图分类号:

X913.4安全系统工程

邹宇航, 吴晓莉, 李默涵. 基于Bayesian-SEM的特种车辆乘员决策行为影响因素研究[J]. 中国安全科学学报, 2026, 36(7): 29-35.

Zou Yuhang, Wu Xiaoli, Li Mohan. Research on influencing factors of decision-making behavior of occupants in special vehicles based on Bayesian-SEM[J]. China Safety Science Journal, 2026, 36(7): 29-35.

图/表 11

图1

表1

火控任务下特种车辆乘员决策因素特性

类别	决策影响因素	影响因素与任务关联	影响因素作用特性
外源性因素	时间压力	火控任务中,从目标识别到火力分配的决策窗口极短,时间限制造成任务困难	1.迅速识别和跟踪目标,生成火力计划; 2.需要快速响应和调整发射参数
	信息动态性	乘员执行任务过程中获取到的信息不断变化,包括敌我态势等;乘员所处环境,如天气、行驶环境等	1.战场信息易变需要乘员实时调整发射条件; 2.复杂地形、天气变化等增加规划难度
	空间限制	车辆内空间狭小,观察窗视野受限,影响乘员操作^[14]	受限的操作空间影响射手的操控灵活性
	任务复杂性	任务中的信息量庞大且复杂,尤其是在多任务并行的情况下,任务复杂性增大可能导致认知负荷过重^[11]	1.多目标信息源增加乘员脑力资源分配难度; 2.多变量的综合考虑增加短时间内规划难度
内源性因素	任务熟悉度	对特定战场地形或标准交战流程等的熟悉程度,影响信息处理模式与决策水平^[15]	高任务熟悉度能更有效地分配火力资源,将个人脑力资源合理分配到对车辆系统参数把控
	知识经验水平	高水平的知识与经验使乘员能够凭借直觉和经验快速做出决策,减少分析时间^[16]	个人经验丰富能做出更合理的火力分配决策,有助于提高命中率和操控精度
	工作记忆	火控任务中需暂时记住多个目标的参数、上级指令的变更等,存在于乘员进行信息处理过程^[17]	1.与处理任务变量和技术参数有关; 2.与乘员个人情绪控制等表现相关
	认知负荷	个人认知资源是有限的,遵循“总量不变,此多彼少”的原则^[18]	增加的认知负荷使注意力分散,提高忽略关键信息的概率;或使得乘员反应行为变得迟缓
	心理应激	不同的心理应激状态会诱发不同程度的情绪变化^[19]	不同的情绪状态可能对乘员火力资源分配、弹道规划等决策行为准确性产生影响

表1

表2

图2

表3

表4

图3

图4

表5

表6

图5

参考文献 22

[1]	王洁, 方卫宁, 李广燕. 基于多资源理论的脑力负荷评价方法[J]. 北京交通大学学报, 2010, 34(6):107-110.
	Wang Jie, Fang Weining, Li Guangyan. Mental workload evaluation method based on multi-resource theory model[J]. Journal of Beijing Jiaotong University, 2010, 34(6):107-110.
[2]	程阁, 刘航. 基于改进模糊AHP法的指挥员作战决策行为效果评估[J]. 舰船电子工程, 2011, 31(5):51-53, 61.
	Cheng Ge, Liu Hang. Evaluation of the commander's combat decision-making behavior impact based on fuzzy-AHP[J]. Ship Electronic Engineering, 2011, 31(5):51-53, 61.
[3]	毛明, 马士奔, 黄诗喆. 主战坦克火力、机动和防护性能与主要总体尺寸的关系研究[J]. 兵工学报, 2017, 38(7):1443-1450. doi: 10.3969/j.issn.1000-1093.2017.07.024
	Mao Ming, Ma Shiben, Huang Shizhe. Research on relationship among firepower, mobility, protection performance and general dimensions of main battle tank[J]. Acta Armamentarii, 2017, 38(7): 1443-1450. doi: 10.3969/j.issn.1000-1093.2017.07.024
[4]	刘维平, 渠景阳, 王锋, 等. 基于SEM的装甲车辆驾驶作业疲劳主要影响因素识别研究[J]. 装甲兵工程学院学报, 2012, 26(2):39-42.
	Liu Weiping, Qu Jingyang, Wang Feng, et al. Research on the major influencing factors identification of armored vehicle driving work fatigue based on SEM[J]. Journal of Academy of Armored Force Engineering, 2012, 26(2):39-42.
[5]	张磊, 彭金栓, 陈晓利. 驾驶分心行为量表设计与影响因素分析[J]. 中国安全科学学报, 2021, 31(11):39-46. doi: 10.16265/j.cnki.issn 1003-3033.2021.11.006
	Zhang Lei, Peng Jinshuan, Chen Xiaoli. Design of driving distraction behavior scale and analysis of influencing factors[J]. China Safety Science Journal, 2021, 31(11):39-46. doi: 10.16265/j.cnki.issn 1003-3033.2021.11.006
[6]	周孙夏, 刘双, 完颜笑如, 等. 噪声环境下告警冗余设计模式对装甲车乘员响应操作绩效的影响[J]. 北京航空航天大学学报, 2026, 52(1):286-293.
	Zhou Sunxia, Liu Shuang, Wanyan Xiaoru, et al. Influence of alarm redundancy design on crew response performance of armored vehicles under noise environment[J]. Journal of Beijing University of Aeronautics and Astronautics, 2026, 52(1):286-293.
[7]	郭司南, 完颜笑如, 刘双, 等. 智能化设计与信息加工通道复杂度对装甲车乘员脑力负荷的影响[J]. 兵工学报, 2021, 42(2):234-241. doi: 10.3969/j.issn.1000-1093.2021.02.002
	Guo Sinan, Wanyan Xiaoru, Liu Shuang, et al. Influences of intelligent design and information processing modality complexity on occupant mental workload[J]. Acta Armamentarii, 2021, 42 (2): 234-241. doi: 10.3969/j.issn.1000-1093.2021.02.002
[8]	范丁元. 基于贝叶斯结构方程模型的城际铁路客运需求影响因素分析[J]. 物流技术, 2021, 40(2):48-52.
	Fan Dingyuan. Analysis of influencing factors of passenger transport demand for intercity rail-transit based on Bayesian SEM[J]. Logistics Technology, 2021, 40(2): 48-52.
[9]	刘维平, 聂俊峰, 刘西侠. 基于多资源理论的乘员信息处理作业认知行为建模研究[J]. 兵工学报, 2017, 38(6):1215-1222. doi: 10.3969/j.issn.1000-1093.2017.06.022
	Liu Weiping, Nie Junfeng, Liu Xixia. Cognitive modeling for crew's information processing based on multi-resource theory[J]. Acta Armamentarii, 2017, 38(6):1215-1222. doi: 10.3969/j.issn.1000-1093.2017.06.022
[10]	Cheong Y, Shehab R L, Ling Chen. Effects of age and psychomotor ability on kinematics of mouse-mediated aiming movement[J]. Ergonomics, 2013, 56(6):1006-1020. doi: 10.1080/00140139.2013.781682 pmid: 23586659
[11]	Wickens C D. Multiple resources and mental workload[J]. The Journal of the Human Factors and Ergonomics Society, 2008, 50(3):449-455. doi: 10.1518/001872008X288394
[12]	马亚龙, 冯军星, 董珊珊. 坦克分队作战指挥决策仿真方法研究[J]. 系统仿真学报, 2013, 25(增刊1):244-248.
	Ma Yalong, Feng Junxing, Dong Shanshan. Simulation method study of combat command decision-making for tank-unit[J]. Journal of System Simulation, 2013, 25(S1): 244-248.
[13]	刘东阳, 朱连章. 基于多Agent系统的坦克分队作战仿真与建模[J]. 电脑知识与技术, 2011, 7(32):7983-7985.
	Liu Dongyang, Zhu Lianzhang. Modeling and simulation of tank unit combat based on multi-agent system[J]. Computer Knowledge and Technology, 2011, 7(32): 7983-7985.
[14]	Pang Liping, Zhang Jie, Wanyan Xiaoru, et al. Field study of neutrality cabin temperature for Chinese passenger in economy class of civil aircraft[J]. Journal of Thermal Biology, 2018, 78:312-319. doi: S0306-4565(18)30266-3 pmid: 30509653
[15]	Yanko M R, Spalek T M. Route familiarity breeds inattention: a driving simulator study[J]. Accident Analysis & Prevention, 2013, 57:80-86. doi: 10.1016/j.aap.2013.04.003
[16]	傅斌贺, 刘维平, 刘西侠, 等. 基于认知能力的乘员信息作业行为差异致因分析[J]. 航天医学与医学工程, 2019, 32(1):19-24.
	Fu Binhe, Liu Weiping, Liu Xixia, et al. Causation analysis of difference in crew information operation behavior based on cognitive ability[J]. Space Medicine & Medical Engineering, 2019, 32(1): 19-24.
[17]	袁伟, 张会明, 赵天欣, 等. 考虑工作记忆容量和时间压力影响的驾驶绩效分析[J]. 汽车工程, 2024, 46(5):862-873.
	Yuan Wei, Zhang Huiming, Zhao Tianxin, et al. Driving performance analysis considering the effect of working memory capacity and time pressure[J]. Automotive Engineering, 2024, 46(5):862-873.
[18]	Paas F G W C, Van Merriënboer J J G. Instructional control of cognitive load in the training of complex cognitive tasks[J]. Educational Psychology Review, 1994, 6(4):351-371. doi: 10.1007/BF02213420
[19]	张舒, 杨齐玉, 余欣柔, 等. 情绪对低感觉寻求者安全行为决策的影响: 来自ERPs的证据[J]. 中国安全科学学报, 2024, 34(8):43-52. doi: 10.16265/j.cnki.issn1003-3033.2024.08.1213
	Zhang Shu, Yang Qiyu, Yu Xinrou, et al. Effects of emotion on safety behavior decision-making of low sensation seekers: evidence from ERPs[J]. China Safety Science Journal, 2024, 34(8): 43-52. doi: 10.16265/j.cnki.issn1003-3033.2024.08.1213
[20]	李辉, 景国勋, 贾智伟, 等. 驾驶员注意力分配定量方法研究[J]. 中国安全科学学报, 2009, 19(2):148-151.
	Li Hui, Jing Guoxun, Jia Zhiwei, et al. Quantitative measuring method for the allocation of drivers' attention[J]. China Safety Science Journal, 2009, 19(2):148-151.
[21]	刘鹏, 吕曦, 李志忠. 任务复杂度对自动化意识的影响[J]. 航空学报, 2015, 36(11):3678-3686. doi: 10.7527/S1000-6893.2015.0166
	Liu Peng, Lyu Xi, Li Zhizhong. Effect of task complexity on automation awareness[J]. Acta Aeronautica et Astronautica Sinica, 2015, 36(11):3678-3686. doi: 10.7527/S1000-6893.2015.0166
[22]	劳可夫. 基于多群组结构方程模型的绿色价值结构研究[J]. 中国人口·资源与环境, 2012, 22 (7):78-84.
	Lao Kefu. Research on green value structure based on multi-group structural equation modeling[J]. China Popuation, Resources and Environment, 2012, 22 (7):78-84.

影响因素	目标感知	态势判定	火力决控	综合权重
阶段权重	0. 374 2	0. 341 6	0. 284 2	—
时间压力	0. 383 3	0. 249 5	0. 367 2	0. 103 0
信息动态性	0. 459 9	0. 223 6	0. 316 5	0. 128 8
空间限制	0. 290 1	0. 433 9	0. 276 0	0. 072 5
任务复杂性	0. 332 2	0. 244 9	0. 422 9	0. 262 3
任务熟悉度	0. 253 3	0. 266 2	0. 480 5	0. 070 0
知识经验水平	0. 268 6	0. 247 0	0. 484 5	0. 069 8
工作记忆	0. 256 8	0. 264 7	0. 478 6	0. 038 6
认知负荷	0. 375 8	0. 339 3	0. 284 8	0. 099 3
心理应激	0. 370 6	0. 318 7	0. 310 7	0. 155 6

潜变量	观测变量含义	表示方法
心理应激${\xi }_{1}$	乘员执行任务时处于疲劳状态S₁	y₁
	乘员处于睡眠不足状态S₂	y₂
	乘员操作车辆系统出现机械性缓慢状态S₃	y₃
时间压力${\xi }_{2}$	乘员获取到的信息不足以在短时间内完成任务T₁	y₄
	乘员完成任务所用时间过长T₂	y₅
	乘员短时间对任务信息做出错误判断T₃	y₆
任务复杂度${\xi }_{3}$	执行当前任务同时需要完成次任务(如弹药选择、车辆驾驶等)C₁	y₇
	当前任务需拆解成细小的多个任务叠加处理C₂	y₈
	乘员在杂乱环境增加任务处理难度C₃	y₉
信息动态变化${\xi }_{4}$	乘员在当前任务时刚经历失败任务I₁	y₁₀
信息动态变化${\xi }_{4}$	乘员未能在短时间内实现有效的多源信息验证I₂	y₁₁
决策水平${\xi }_{5}$	乘员仓促做出错误决策行为D₁	y₁₂
决策水平${\xi }_{5}$	乘员未能正确处理车辆技术任务D₂	y₁₃

潜变量	心理应激	时间压力	任务复杂度	信息动态性	决策水平
心理应激	0.83	—	—	—	—
时间压力	0.58^**	0.79	—	—	—
任务复杂度	0.42^**	0.61^**	0.81	—	—
信息动态性	0.52^**	0.64^**	0.48^**	0.76	—
决策水平	-0.37^**	-0.52^**	-0.46^**	-0.41^**	0.78

模型	χ²	DF	Δχ²(ΔDF)	p值	结论
形态相同	—	—	—	—	—
测量权重相等	2.884	8	2.884 (8)	0.941	接受不变性
结构权重相等	22.286	21	19.403 (13)	0.111	接受不变性

路径	Bayesian估计结果		ML估计结果
路径	Mean	S.E.	Estimate	S.E.
心理应激← 信息动态变化	1.80^*	0.01	2.06^*	0.39
决策水平← 心理应激	0.96^*	0.01	1.47^*	0.40
时间压力← 任务复杂度	-0.33^*	0.01	-0.28^*	0.11
决策水平← 时间压力	-0.50^*	0.02	-1.28^*	0.61
时间压力← 信息动态变化	1.64^*	0.02	1.70^*	0.39