多模态人因风险评估驱动的访问控制模型OA
Multimodal Human-factor Risk Assessment-driven Access Control Model
在生产场景中,人员管理是安全管理中的重要环节,而情绪往往是影响人员表现的重要因素.本文提出一种以多模态人因风险评估结果为驱动的智能访问控制模型,该模型通过实时监测操作人员情绪状态,并以此为依据,进行风险等级评估,做到提前预警,动态调整其操作权限,保证系统的安全.首先,本文构建多模态情绪识别模块,融合语音情绪与面部情绪识别结果,弥补单模态存在的信息不足问题,对操作人员的情绪状态进行实时监测.其次,本文采用风险矩阵法,对风险可能性和后果严重程度进行量化,以多模态情绪识别结果作为依据,实时评估人因风险等级.最后,本文将人因风险评估等级作为决策依据,对操作人员权限进行实时调整,限制高风险等级下的敏感行为.本文模型将操作人员的情绪状态与系统操作权限进行了关联,从而实现了人因安全管理的被动应对到主动调整的转换.
In production environments,personnel management is a critical component of security management,and emotion is often a significant factor influencing personnel performance.This paper proposes an intelligent access control model driven by the results of a multi-modal human factors risk assessment.The model monitors the emotional state of operators in real-time and uses it as a basis to assess risk levels to provide early warnings,dynamically adjust operational permissions,and thereby ensure system security.Firstly,this paper constructs a multi-modal emotion recognition module that fuses results from both vocal and fa-cial emotion recognition to compensate for the informational deficiencies of single-modality methods and monitor the emotional state of the operator in real time.Secondly,this paper adopts the risk matrix method to quantify the possibility of risks and the se-verity of their consequences,and uses the results of multimodal emotion recognition as the basis to assess the human factor risk level in real time.Finally,this paper uses the assessed human factors risk level as a decision-making basis to adjust operator per-missions in real-time,restricting sensitive behaviors under high-risk conditions.The proposed model correlates the emotional state of operators with their system operation permissions,thereby achieving transformation in human factors security manage-ment from reactive response to proactive adjustment.
李涛;董钊辰;刘忻
兰州大学信息科学与工程学院,甘肃 兰州 730000兰州大学信息科学与工程学院,甘肃 兰州 730000兰州大学信息科学与工程学院,甘肃 兰州 730000
信息技术与安全科学
风险评估访问控制情绪识别机器学习
risk assessmentaccess controlemotion recognitionmachine learning
《计算机与现代化》 2026 (7)
33-38,51,7
甘肃省重点研发计划-工业类(23YFGA0010)
评论