基于声纹识别的电网调度认证系统设计

Design of Power Grid Dispatching Authentication System Based on Voiceprint Recognition

随着我国电网智能化规模日益扩大,如何保障发布安全可靠的电网调度指令成为研究热点。但现有研究中大多缺乏调度员身份可靠校验,存在一定的安全隐患。针对该安全问题,提出构建基于声纹识别的调度认证系统,由数据层、功能层、接口层和业务应用层4层架构组成,采用卷积神经网络从短时语音中提取声纹身份特征,融合基于动态口令的身份认证方法,在短时语音识别和欺骗性语音检测方面取得良好的效果,解决电网调度场景中可能存在的多种未知欺骗攻击和错误调度指令问题,并对调度指令进行正确性校验,避免错误指令。基于声纹识别的电网调度系统可高效地响应用户请求,在执行身份认证的同时给予及时有效的相关反馈,提高调度工作的安全性和质量,具备实际应用价值。

With the increasing scale of intelligence in China’s power grid,how to guarantee the issuance of safety and reliability of power grid dispatching commands has become a research hotspot.However,most of the existing research lacks reliable verification of dispatcher identity,which poses certain security risks.Aiming at the security issues,the dispatching authentication system based on voiceprint recognition is proposed,which consists of a four-layer architecture of data layer,functional layer,interface layer,and business application layer.Convolutional neural networks are used to extract voiceprint identity features from short-term speech,and the identity authentication methods based on dynamic password are integrated.It achieves good results in short speech recognition and deceptive speech detection,solves various unknown spoofing attacks and incorrect dispatching instructions that may exist in power grid scheduling scenarios,and verifies the correctness of dispatching instructions to avoid incorrect instructions.The power grid dispatching authentication system based on voiceprint recognition can efficiently respond to user requests,provide timely and effective feedback while performing identity authentication,improve the security and quality of dispatching work,and have practical application value.