摘要
数字孪生业务的差异化需求,要求孪生层能够快速可靠地获取物理系统的网络状态和终端设备的感知信息等数据,且具备防御敌意干扰和通信干扰的能力。智能无线通信抗干扰技术采用强化学习优化发射功率和信道选择等通信策略,提升抗干扰性能。然而,目前大多方法假设能够及时准确地获取通信质量反馈和具备良好的干扰估计能力,在数字孪生业务中依然面临抗干扰优化速度慢和承载业务差异化需求难以满足的问题。因此,提出了一种面向数字孪生的智能无线通信抗干扰方案,挖掘干扰特征以及数字孪生业务的服务质量需求,引入关于时延的安全约束并设计效益评估方法,优化终端设备的通信抗干扰策略。在基于传感手套和机械臂的数字孪生手势重建系统中,实验验证了所提方案与现有智能通信抗干扰方案相比,降低了时延和通信能耗,提高了手势重建成功率。
The diverse requirements of digital twin services require the twin layer to quickly and reliably obtain the data,such as network states of physical systems and sensing information of terminal devices,and the ability against the hostile jamming and communication interference is also needed.Intelligent wireless anti-jamming communication techniques apply reinforcement learning to optimize the communication policies such as transmit power and channel selection to improve anti-jamming performance.However,most existing methods assume to have accurate and real-time feedback of the communication quality and good jamming estimation capabilities,and they still face the problems of slow anti-jamming optimization and fail to satisfy diverse requirements of bearing services.Therefore,this paper proposes an intelligent anti-jamming wireless communication scheme for digital twins,which exploits jamming characteristics and quality of service requirements of digital twin services,introduces a delay-based security constraint,and designs the utility evaluation method to optimize the anti-jamming policy of terminal devices.In a digital twin gesture reconstruction system based on the sensing glove and robotic arm,experiment results show that the proposed scheme reduces the delay and communication energy and improves the successful rate of gesture reconstruction compared with existing intelligent anti-jamming schemes.
作者
肖亮
陈顺
王烁华
吕泽芳
杨和林
XIAO Liang;CHEN Shun;WANG Shuohua;LV Zefang;YANG Hein(School of Informatics,Xiamen University,Xiamen 361005,China)
出处
《移动通信》
2023年第6期8-13,共6页
Mobile Communications
基金
国家自然科学基金区域联合重点项目(U21A20444)
国家自然科学基金面上项目(61971366)。
关键词
数字孪生
无线通信
敌意干扰
强化学习
digital twin
wireless communications
hostile jamming
reinforcement learning
作者简介
肖亮:博士毕业于美国罗格斯大学,现任厦门大学信息学院教授、博士生导师;陈顺:厦门大学在读硕士研究生,主要研究方向为智能无线通信和强化学习;王烁华:厦门大学在读硕士研究生,主要研究方向为智能无线通信和强化学习;吕泽芳:厦门大学在读博士研究生,主要研究方向为无线通信和强化学习;杨和林:博士毕业于新加坡南洋理工大学,现任厦门大学信息学院副教授、博士生导师,从事无线通信、智能通信和机器学习等方向的研究。
