绿色认知无线传感器网络的分布式多目标联合能效优化

This study proposes a distributed multi-objective optimization scheme combining distributed technology to optimize the energy efficiency of wireless sensor networks. Via this scheme the energy consumption of spectrum detection and information perception is balanced. Moreover, the performance of spectral availability is improved along with the efficiency of application oriented information perception. 1) The technology of distributed cooperative spectrum sensing is studied for energy-constrained cognitive radio sensor network nodes. By leveraging the distributed gradient-descent reinforcement learning method, the mechanism of nodes sleep scheduling is proposed to guarantee the spectrum detection performance. 2) As the uncertainty of channel state information, a partially-observable stochastic game theory is developed to obtain optimal channel and power allocation results. 3) The application requirements of cognitive radio sensor network are studied. According to the environmental signal source distribution characteristics, a distributed water-filling algorithm is proposed to joint optimize the energy efficiency of precise spectrum sensing and data transmission, by organic combining the nodes sleep scheduling mechanism with the interference constraints. This project develops an effective distributed multi-objective optimization scheme for the energy equilibrium of green cognitive radio sensor networks, and a parallel distributed computing architecture can be obtained with strong extensibility also. Through the research, the green communication technologies for cognitive radio sensor networks will be promoted.

本项目针对绿色认知无线传感器网络的能效优化问题，提出一种分布式多目标联合优化方法，均衡频谱检测和信息感知的能量消耗，提高网络的频谱利用率和面向应用的信息感知性能。1）研究能量受限的认知传感器网络节点的分布式协作频谱检测技术，设计分布式次梯度学习算法，提出保证频谱检测性能的节点休眠调度机制；2）对认知节点获得的不确定信道状态信息，设计部分可观测随机博弈算法对各个可用频段的发送功率进行控制，实现最佳信道分配和功率控制；3）研究认知无线传感器网络的应用需求，根据环境信号源分布特性，将节点休眠调度机制和认知场景下的干扰约束有机融合，设计分布式认知注水算法，实现保障应用数据准确率的频谱认知与数据感知的能效联合优化。通过本项目的研究，将为绿色认知传感器网络的能量均衡使用提出一种有效的并行分布式多目标联合优化方法,得到可扩展性强的分布式计算架构，对认知无线传感器网络的绿色通信技术发展起到一定的促进作用。

本项目针对绿色认知无线传感器网络的能效优化问题，提出一种分布式多目标联合优化方法，均衡频谱检测和信息感知的能量消耗，提高网络的频谱利用率和面向应用的信息感知性能。通过本项目的研究，将为绿色认知传感器网络的能量均衡使用提出一种有效的并行分布式多目标联合优化方法,得到可扩展性强的分布式计算架构，对认知无线传感器网络的绿色通信技术发展起到一定的促进作用。. 申请书中设定的工作目标全部完成，包括：设计分布式次梯度学习算法，研究保证频谱检测性能的节点休眠调度机制、设计最佳信道分配和功率控制的部分可观测随机博弈算法对各个可用频段的发送功率进行控制、设计保障应用数据准确率的频谱认知与数据感知的能效联合优化的分布式认知注水算法。. 受本项目支持，相关的研究成果已发表在国内外权威学术刊物和国际顶级的计算机网络、通信和控制领域的国际会议GLOBECOM、ICC、ACC、ECCE上。发表学术论文41篇，其中SCI收录论文13篇，EI收录论文25篇，CSSCI论文1篇。已获授权发明专利2项，受理发明专利5项，授权软件著作权16项。获得奖励12项。培养毕业博士研究生2名，硕士研究生12名。.