认知无线可充电传感网的安全资源分配研究

To solve the problems of spectrum scarcity and energy constraint in wireless sensor networks, this project employs cognitive radio and energy self-supply (energy harvesting/wireless energy transfer) techniques to formulate cognitive wireless rechargeable sensor network (CWRSN). However, to realize the green and secure communication of CWRSN in licensed channels, we are facing the following key problems: inaccuracy for modeling resources due to their dynamics and variability, unsuitability for complex encryption and decryption algorithms due to serious resource constraints, and unavailability for allocating resources due to their coupling. To solve these problems, this project first establishes multi-dimensional models for temporal, spatial and spectrum resources; then, employs physical layer security techniques including collaborative relay and cooperative jamming to design light-weight secure communication mechanism; finally, adopts cross-layer design and joint optimization methods to allocate resources coupled in different protocol layers, and proposes secure resource allocation algorithms. Furthermore, a prototype is developed, with which simulations and experiments are implemented to verify the validity of the proposed methods. The success of this project will enhance spectrum efficiency, energy efficiency and communication security simultaneously, and can serve as a solid theoretical foundation for establishing high reliability, high energy efficiency and high security CWRSN.

为解决无线传感网面临的频谱稀缺和能量受限问题，本项目应用认知无线电和能量自供给技术（能量收集/无线充电）建立了认知无线可充电传感网。然而，实现其在授权信道的绿色和安全通信还面临以下关键问题：资源动态多变难以准确建模，资源严重受限难以支持复杂加密解密算法，以及资源复杂耦合难以有效分配。为解决上述问题，本项目首先对多维时空频资源进行建模；然后，采用合作中继传输和协作干扰监听等物理层安全技术设计轻量级安全通信机制；最后，采用跨层设计与联合优化的方法对复杂耦合在不同协议层的资源进行联合分配，设计安全资源分配算法。在此基础上，设计并开发原型系统，通过物理实验和仿真分析相结合的方式验证所提方法的有效性。本项目的成功开展将同时提高频谱效率、能量效率和通信安全性，为设计高可靠、高能效和高安全的认知无线可充电传感网提供坚实的理论基础。

认知无线电和无线充电技术的快速发展为解决无线传感网的频谱稀缺和能量受限问题提供了全新的解决途径。本项目通过构建认知无线可充电传感网，并从频谱认知、资源分配、传输调度、安全机制等方面开展深入研究，构建了认知无线传感网资源分配的理论方法体系。在频谱感知方面，提出了异构认知无线传感网的多频带协作频谱感知方法，发明了基于机器学习的高效快速频谱认知方法，实现了高能效的频谱认知；在资源分配方面，基于深度强化学习（DRL）方法，提出了基于单智能体DRL的通信资源分配方法、基于多智能体DRL的通信与计算资源联合分配方法，实现了大规模异构任务的高并发通信；在传输调度方面，推导并得到了多跳认知无线可充电传感网的端到端中断概率的准确表达式和近似表达式，提出了基于粒子群优化算法的动态资源分配方法，实现了传输时间和发射功率的近似最优分配，确保中断概率的最小化；在安全机制方面，制定了国家军用标准。为验证项目成果的有效性，基于软件无线电USRP，搭建了大规模MIMO天线阵列系统，有效支持了本项目的频谱感知、资源分配与调度、协议设计等工作。基于上述成果，项目组提出并建立了面向工业现场控制的工业5G网络体系WIA-NR。.项目组累计发表或录用学术论文15篇，其中SCI检索9篇，EI检索12篇；申请国际PCT专利1项，国家发明专利9项，已授权1项；形成《无线同步自组织传感网》系列国家军用标准GJB 9605-2019，4项，已发布实施。部分成果得到《中国科学报》《中科院之声》等媒体报道。

内容获取失败，请点击重试