基于VR试验的地下综合体人员疏散决策OA北大核心CSTPCD
Pedestrian evacuation decision-making for underground complexes based on VR experiments
为了探究人员在疏散过程中的出口选择等决策行为规律,预测疏散人员对疏散出口的选择情况,应用虚拟现实(VR)技术搭建地下空间综合体场景,并开展人员疏散试验.首先,借助SPSS软件对试验结果进行统计分析,构建无序多分类Logistic回归模型,识别主要影响因素;然后,计算各类影响因素的贡献并建立相关性分析模型;最后,基于支持向量机(support vector machine,SVM)算法,建立人员疏散决策模型.结果表明:疏散人员倾向于选择初始朝向的出口进行疏散;当面对的多个出口距离相近时,人员与出口的距离对疏散人员出口选择的影响较小;基于SVM的疏散决策模型的预测准确率为93％.
In order to explore the decision behavior rules such as exit selection in the evacuation process,and predict the selection of evacuation exits by evacuees,we used virtual reality(VR)technology to build un-derground space complex scenes and carry out pedestrian evacuation experiments.First,survey results were statistically analyzed using SPSS software,and disordered multi-classification Logistic regression models were built to identify the main influencing factors.Then,we calculate the contribution of various influen-cing factors and established a correlation analysis model.Finally,based on the support vector machine(SVM)algorithm,a pedestrian evacuation decision-making model was established to predict the exit choice of pedestrians in the evacuation.The results show that pedestrians tend to choose the initially orientated exits for evacuation.In addition,when facing multiple exits with similar distances,the distance from the ex-it has little impact on pedestrian exit choice.The prediction accuracy of the evacuation decision-making model based on SVM is 93％.
余翔宇;王维莉
上海海事大学物流科学与工程研究院,上海 201306
环境科学
地下综合体人员疏散决策行为虚拟现实(VR)技术Logistic回归分析支持向量机
underground complexpedestrian evacuation decision-making behaviorvirtual reality(VR)tech-nologyLogistic regression analysissupport vector machine(SVM)
《安全与环境工程》 2024 (004)
125-132 / 8
国家自然科学基金项目(71904116)
评论