大规模相控阵是解决毫米波无线传输距离受限的核心关键技术.传统的毫米波相控阵通常基于化合物半导体芯片加以实现,该类芯片成本高昂且难以实现系统单片集成,极大地限制了传统相控阵的应用范围.本文报道了基于CMOS成熟工艺的毫米波芯片...大规模相控阵是解决毫米波无线传输距离受限的核心关键技术.传统的毫米波相控阵通常基于化合物半导体芯片加以实现,该类芯片成本高昂且难以实现系统单片集成,极大地限制了传统相控阵的应用范围.本文报道了基于CMOS成熟工艺的毫米波芯片设计及收发通道数为4096(4096发射/4096接收)的超大规模集成相控阵实现技术.CMOS体硅工艺具有集成度高、成本低廉等优势,但面临有源器件高频性能差、无源器件及互连线高频损耗大、高低温性能差异大等一系列技术瓶颈.通过引入电流复用跨导增强型低噪声放大器、基于新型版图结构的高效率功率放大器、矢量调制型数控无源移相器、基于电容补偿的超宽带衰减器、紧凑型功分器,以及高低温自适应偏置电路等技术,可以较好地解决CMOS体硅工艺所面临的上述瓶颈问题.基于65 nm CMOS体硅工艺,所实现的Ka频段CMOS相控阵芯片噪声系数为3.0 d B,发射通道效率为15%,无需校准即可实现精确幅相控制,相关测试结果表明所研制的低成本相控阵芯片具有集成度高、幅相控制精确等优势,噪声系数等关键技术指标接近砷化镓工艺.以此为基础,本文给出了基于多层混压PCB工艺的1024发射/1024接收超大规模"集成相控阵"设计技术,并将其扩展至4096发射/4096接收相控阵规模,最后给出了低成本、高集成宽带卫星移动通信终端在车载和船载条件下的示范应用结果.展开更多
In this paper,we investigate the effective deployment of millimeter wave(mmWave)in unmanned aerial vehicle(UAV)-enabled wireless powered communication network(WPCN).In particular,a novel framework for optimizing the p...In this paper,we investigate the effective deployment of millimeter wave(mmWave)in unmanned aerial vehicle(UAV)-enabled wireless powered communication network(WPCN).In particular,a novel framework for optimizing the performance of such UAV-enabled WPCN in terms of system throughput is proposed.In the considered model,multiple UAVs monitor in the air along the scheduled flight trajectory and transmit monitoring data to micro base stations(mBSs)with the harvested energy via mmWave.In this case,we propose an algorithm for jointly optimizing transmit power and energy transfer time.To solve the non-convex optimization problem with tightly coupled variables,we decouple the problem into more tractable subproblems.By leveraging successive convex approximation(SCA)and block coordinate descent techniques,the optimal solution is obtained by designing a two-stage joint iteration optimization algorithm.Simulation results show that the proposed algorithm with joint transmit power and energy transfer time optimization achieves significant performance gains over Q-learning method and other benchmark schemes.展开更多
文摘大规模相控阵是解决毫米波无线传输距离受限的核心关键技术.传统的毫米波相控阵通常基于化合物半导体芯片加以实现,该类芯片成本高昂且难以实现系统单片集成,极大地限制了传统相控阵的应用范围.本文报道了基于CMOS成熟工艺的毫米波芯片设计及收发通道数为4096(4096发射/4096接收)的超大规模集成相控阵实现技术.CMOS体硅工艺具有集成度高、成本低廉等优势,但面临有源器件高频性能差、无源器件及互连线高频损耗大、高低温性能差异大等一系列技术瓶颈.通过引入电流复用跨导增强型低噪声放大器、基于新型版图结构的高效率功率放大器、矢量调制型数控无源移相器、基于电容补偿的超宽带衰减器、紧凑型功分器,以及高低温自适应偏置电路等技术,可以较好地解决CMOS体硅工艺所面临的上述瓶颈问题.基于65 nm CMOS体硅工艺,所实现的Ka频段CMOS相控阵芯片噪声系数为3.0 d B,发射通道效率为15%,无需校准即可实现精确幅相控制,相关测试结果表明所研制的低成本相控阵芯片具有集成度高、幅相控制精确等优势,噪声系数等关键技术指标接近砷化镓工艺.以此为基础,本文给出了基于多层混压PCB工艺的1024发射/1024接收超大规模"集成相控阵"设计技术,并将其扩展至4096发射/4096接收相控阵规模,最后给出了低成本、高集成宽带卫星移动通信终端在车载和船载条件下的示范应用结果.
文摘混合波束赋形在射频模拟域波束赋形的基础上,通过在基带实现数字域波束赋形来进一步提高数据速率.本文针对基于混合波束赋形的室内毫米波MIMO(Multiple Input Multiple Output)系统进行性能分析,分别评估了迫零(Zero Force,ZF)接收机、最小均方误差(Minimum Mean Square Error,MMSE)接收机以及匹配滤波(Matched Filter,MF)接收机在基于码本的最优波束选择下的传输速率,并对接收机在不同的信道环境下进行性能评估.仿真结果表明,无论视距(Line-Of-Sight,LOS)传播路径是否被遮挡,MMSE接收机和ZF接收机在总发射功率较高时均具有接近容量的性能;在总发射功率较低时MMSE接收机的性能则略优于ZF接收机.相反,MF接收机始终保持较差的性能.系统可根据总发射功率大小以及接收机的实现复杂度来选取合适的接收机.本文所研究的系统可应用于实际室内无线个域网(Wireless Personal Area Networks,WPAN)以实现高速短距无线通信.
文摘In this paper,we investigate the effective deployment of millimeter wave(mmWave)in unmanned aerial vehicle(UAV)-enabled wireless powered communication network(WPCN).In particular,a novel framework for optimizing the performance of such UAV-enabled WPCN in terms of system throughput is proposed.In the considered model,multiple UAVs monitor in the air along the scheduled flight trajectory and transmit monitoring data to micro base stations(mBSs)with the harvested energy via mmWave.In this case,we propose an algorithm for jointly optimizing transmit power and energy transfer time.To solve the non-convex optimization problem with tightly coupled variables,we decouple the problem into more tractable subproblems.By leveraging successive convex approximation(SCA)and block coordinate descent techniques,the optimal solution is obtained by designing a two-stage joint iteration optimization algorithm.Simulation results show that the proposed algorithm with joint transmit power and energy transfer time optimization achieves significant performance gains over Q-learning method and other benchmark schemes.
