基于CSI商联合AOA的增强型指纹定位方法

针对当前指纹定位技术需要基于多个接入点(AP)被动定位,导致单AP室内场景使用受限的问题,提出一种信道状态信息(CSI)商联合到达角(AOA)的室内指纹定位方法:指出技术关键是利用单链路上的多维信号参数来构建指纹模型;利用多进多出(MIMO)系统空间分集,构建CSI商以获得更稳健的CSI指纹信号,并结合多信号分类(MUSIC)算法原理设计一种多载波AOA指纹表示方法,该方法与原始CSI指纹相比具有区分性;然后为了解决单AP的AOA指纹对称性问题,将CSI商与AOA指纹相结合,得到新的指纹,通过机器学习的方法进行目标位置匹配。实验结果表明,本方法在空教室和实验室环境下的定位准确率分别达到98.96%和97.08%,平均定位误差分别为0.46 m和0.68 m,具有较高的定位性能。

一种融合5G CSI和地磁的集成学习定位方法

针对深度学习算法在多传感器融合定位中容易出现的局部收敛、异质融合性能不佳等问题,本文提出了一种集成双向长短期记忆网络(BiLSTM)和注意力机制的多输入卷积神经网络(CNN)的室内定位算法。该算法首先对5G信道状态信息(CSI)和地磁数据分别进行预处理;然后各自基于独立的分支网络进行离线训练,同时提取指纹数据的空间特征和时序特征,追加注意力机制层;最后在全连接层实现了异质传感器数据的融合定位。在会议室和教学楼大厅的试验结果表明,平均定位误差分别为0.95和1.84 m,相比误差反向传播网络(BPNN)分别提高了48.9%和42.7%,定位精度和系统稳定性均大幅提升。

RSSI和PC-CSI加权融合的指纹定位方法

针对基于RSSI和CSI的指纹定位技术易受环境干扰、定位精度较低的问题,提出了一种基于RSSI指纹和相位修正信道状态信息(phase correct based channel state information,PC-CSI)指纹的加权融合指纹定位技术。基于PC-CSI的指纹定位在传统基于CSI幅值的指纹定位基础上增加相位信息对定位结果进行修正,之后对RSSI指纹和PC-CSI指纹的定位结果加权重定位。实验结果表明,提出的加权融合指纹定位算法与基于CSI的主动定位算法相比,平均定位误差(mean position error,MPE)降低了36.2%,能满足室内定位需求。

RIS辅助MIMO NOMA系统中利用统计CSI的下行传输方法

针对可重构智能反射表面(RIS)辅助多输入多输出(MIMO)非正交多址接入(NOMA)下行传输系统,该文提出利用统计信道状态信息(CSI)的基站发送协方差矩阵与RIS相移矩阵设计方法。首先,在莱斯空间相关信道假设下,利用大维随机矩阵理论,推导了RIS辅助MIMO NOMA系统遍历和速率的确定性表达式;然后,在弱用户速率约束与发送功率受限的条件下,通过最大化确定性大系统近似和速率,利用统计CSI,分别设计了强、弱用户的次优发送协方差矩阵和RIS的相移矩阵。仿真结果表明,所推导的近似表达式具有很好的近似效果,所设计的发送协方差矩阵和相移矩阵能显著提升系统的和速率。

基于特征融合的大规模MIMO系统CSI反馈

信道状态信息(Channel State Information,CSI)反馈是大规模多输入多输出(Multiple-Input Multiple-Output,MIMO)系统的一个关键问题。大规模MIMO系统中基站天线数量巨大,CSI反馈出现了反馈开销大、反馈精度低等问题。为了降低反馈开销,提高反馈精度,采用深度学习方法,提出了一种基于特征融合的CSI反馈网络(Feature Fusion Net,FFNet)。利用基于注意力机制的特征融合在编码器中融合不同尺度的CSI特征,并在解码器中使用多通道多分辨率卷积网络以及通道重排,从而高精度地重建压缩后的CSI。仿真结果表明,与几种经典的深度学习CSI反馈方法相比,在室内和室外信道条件下,均具有更高的反馈精度。

基于GMM-DBC的CSI室内定位算法

针对贝叶斯室内定位技术存在定位精度低及时间复杂度较高的问题,提出了一种基于高斯混合模型和密度聚类(GMM-DBC)的信道状态信息(CSI)定位算法.通过对分模型参数的初次估计构建GMM概率分布模型并进行误差计算;引入确定分模型个数(DSM)策略,结合误差计算结果更新GMM模型参数,减小由模型精度引起的定位误差;基于不同参考点的分布特征,判断参考点间紧密程度,将紧密相连的参考点划为一类,减小搜索范围,降低时间复杂度;根据分簇结果,利用改进的贝叶斯概率算法进行权值计算,得到最终定位结果.实验结果表明:所提算法能较好地提高定位精度,降低时间复杂度.

变化信道稀疏度的CSI反馈方法

在大规模多输入多输出(multiple input multiple output,MIMO)系统中,压缩感知(compressed sensing,CS)技术常用于具有稀疏特性的信道状态信息(channel state information,CSI)反馈。针对CS重构时信道稀疏度通常未知的问题,基于深度展开技术提出了一种变化信道稀疏度的CSI反馈方法(a CSI-feedback method for varying channel sparsity,AVCS)。AVCS将信道稀疏度作为训练参数,学习得到通用的网络架构。随着天线数量增大导致信道(矩阵)维度激增,学习网络所得的相互抑制矩阵会呈现二次增长问题,AVCS利用相互抑制矩阵托普利兹(Toeplitz)特性设计了降维卷积网络,解决CSI反馈时的计算复杂度问题。仿真结果表明,所提方法提高了在大规模MIMO系统下CSI重构的适用性,减少了反馈开销且对信道稀疏度具有鲁棒性。

Imperfect Channel State Information of AF and DF Energy Harvesting Cooperative Networks

Wireless information and powered transfer networks(WIPT) has recently been implemented in 5th generation wireless networks. In this paper, we consider half-duplex relaying system in which the energy constrained relay node collects energy via radio frequency(RF) signals from the surrounding resources. Regarding energy harvesting protocol, we propose power time switching-based relaying(PTSR) architecture for both amplify-and-forward(AF) and decode-and-forward(DF). Especially, we reveal the analytical expressions of achievable throughput, ergodic capacity and energy-efficient in case of imperfect channel state information(CSI) for both AF and DF network. Through numerical analysis, we analyse the throughput performance, energy-efficient and ergodic capacity for different parameters, including power splitting ratio and energy harvesting time. Moreover, we also depict the performance comparison between AF and DF network with perfect and imperfect CSI. The results in numerical analysis reveal that the result of AF relaying network is less significant than DF relaying network in the various scenarios.

基于CSI和K-means-SVR的多指纹库室内定位方法

传统的Wi-Fi室内定位方法需要与所有指纹数据库中的指纹数据进行匹配后才能定位,导致人群聚集区域定位效率不高,体验较差。提出一种基于信道状态信息(CSI)、K均值(K-means)聚类算法与支持向量回归(SVR)算法相结合的多指纹库室内定位方法。该方法首先根据CSI的簇分布特点,利用K-means算法对所有定位点内的CSI数据聚类后得到多个簇的CSI数据;然后,基于多个簇分别建立多个指纹库,并将CSI数据分别存入多个指纹库,进而在每个指纹库中分别训练SVR模型用于Wi-Fi定位。相较于传统的支持向量机(SVM)定位方法,所提方法在离线阶段需要的训练样本更少,定位效率更高;在线阶段,该方法既降低了匹配的复杂度,也提高了定位的精度。由于使用了多指纹库,Wi-Fi定位系统可以根据人流量实时调整资源分配策略,提高服务器运行效率和定位服务体验。

基于WiFi CSI的多特征融合的步态识别

基于WiFi信道状态信息(CSI)的步态识别作为一种新型的生物特征识别方案,从步态周期、步速、行走姿势、身体平衡、足底压力等方面的差异进行分析,从而有效识别行人身份。然而,目前大多数基于WiFi CSI的步态识别方案中主要针对人体的运动特征进行识别,鲜见对人体物理形态特征进行分析,这样就导致了特征提取不完整,系统识别率低,鲁棒性差等问题。针对上述情况,提出了一种结合步行时运动特征和相对静止时物理形态特征的多特征融合身份识别方法。该方法利用了一个双流卷积神经网络(CNN)分别提取粗粒度运动特征和细粒度物理形态特征,并将两部分特征融合,进行身份识别。在真实的场景中进行了实验,最高准确率达到97.1%,明显优于现有的识别方案,实验证明了该方法的有效性和通用性。

基于相控阵的WiFi CSI定位系统的性能预测

WiFi作为当前最重要的通信方式之一,基于WiFi信号的室内定位系统最有望在日常生活中得到广泛地部署应用.最新研究表明,当采用WiFi通信过程中获取的信道状态信息(CSI)对目标进行定位时,系统可实现亚米级的定位精度.然而,实验场景下的定位精度受到测试样点位置、WiFi设备布局、天线布局等诸多因素的影响.因为目前仍缺少WiFi CSI定位性能预测方法,WiFi定位系统部署后往往难以获得预期的精度.为此,面向多样化场景提出WiFi CSI定位性能的预测模型.首先,从CSI定位的基本物理模型出发,定义天线对的误差微元函数,并通过对定位空间的分析生成误差微元矩阵以及定位性能热度图;其次,对天线对进行拓展,通过引入多天线融合方法、多设备融合方法构建通用的CSI定位性能预测模型;最后,为了将真实场景地图考虑在内,提出将上述热度图与场景地图相融合的方法,从而实现场景定制化的性能预测.在理论分析的基础上,结合2个不同场景下的实验数据验证了定位性能预测模型有效性.实验结果表明,实际定位精度的变化趋势与理论模型相吻合,通过理论模型分析可将定位精度优化32%–37%.

Network coding with diversity and outdated channel state information

Physical-layer network coding （PNC） has the potential to significantly improve the throughput of wireless networks where the channels can be modeled as additive white Gaussian noise （AWGN） channel. As extending to mul- tiple channels, this technique requires both amplitude and phase compensation at each transmitter and will lead to inef- ficient systems yielding no diversity even with perfect channel state information （CSI）. In order to avoid these limita- tions, we apply network coding with diversity （NCD） to achieve a form of selection diversity and extend NCD to coop- erative multiple access channels in this paper. However, in practical wireless communication systems, the CSI could become outdated due to the difference between the CSI used in the relay selection and data transmission phases. Hence, the selected relay may not be the best one during data transmission phase due to the dynamic change in the wireless channels. Therefore, we first explore the relation between the present and past CSIs. Exploiting this relationship, the NCD scheme with outdated CSI is investigated based on the past CSI. To evaluate the performance of this scheme, an information-theoretic metric, namely the outage capacity, is studied under this condition.

面向室内人员被动感知的CSI幅值平稳性分析方法

室内复杂环境下,多径效应导致无线信号的接收信号强度(RSS)分辨率较低,非视距条件造成室内人员的检测准确率不佳,现有技术对室内静态人员的漏检率与设备成本偏高,且构建系统较复杂。针对以上问题,提出一种仅使用单接入点(AP)的面向室内人员被动感知的信道状态信息(CSI)幅值平稳性分析方法。首先对采集的CSI提取幅值矩阵并进行数据预处理;在此基础上分别计算CSI幅值的全局稳定度与局部稳定度,最终基于双阈值联合判定方法实现室内有无人员的轻量化快速判定。实验结果表明,该方法在视距下对动态人员的被动感知准确率达99.6%,在非视距条件下对静态人员感知准确率仍可达91.87%。该方法突破了需要携带并修改设备、部署传感器和视距等条件限制,在降低成本的同时有效提高了系统的可靠性与易用性。

Multi-Sinusoidal Waveform Shaping for Integrated Data and Energy Transfer in Aging Channels

Integrated data and energy transfer(IDET)is capable of simultaneously delivering on-demand data and energy to low-power Internet of Everything(Io E)devices.We propose a multi-carrier IDET transceiver relying on superposition waveforms consisting of multi-sinusoidal signals for wireless energy transfer(WET)and orthogonal-frequency-divisionmultiplexing(OFDM)signals for wireless data transfer(WDT).The outdated channel state information(CSI)in aging channels is employed by the transmitter to shape IDET waveforms.With the constraints of transmission power and WDT requirement,the amplitudes and phases of the IDET waveform at the transmitter and the power splitter at the receiver are jointly optimised for maximising the average directcurrent(DC)among a limited number of transmission frames with the existence of carrier-frequencyoffset(CFO).For the amplitude optimisation,the original non-convex problem can be transformed into a reversed geometric programming problem,then it can be effectively solved with existing tools.As for the phase optimisation,the artificial bee colony(ABC)algorithm is invoked in order to deal with the nonconvexity.Iteration between the amplitude optimisation and phase optimisation yields our joint design.Numerical results demonstrate the advantage of our joint design for the IDET waveform shaping with the existence of the CFO and the outdated CSI.

Device-Free Through-the-Wall Activity Recognition Using Bi-Directional Long Short-Term Memory and WiFi Channel State Information

Activity recognition plays a key role in health management and security.Traditional approaches are based on vision or wearables,which only work under the line of sight(LOS)or require the targets to carry dedicated devices.As human bodies and their movements have influences on WiFi propagation,this paper proposes the recognition of human activities by analyzing the channel state information(CSI)from the WiFi physical layer.The method requires only the commodity:WiFi transmitters and receivers that can operate through a wall,under LOS and non-line of sight(NLOS),while the targets are not required to carry dedicated devices.After collecting CSI,the discrete wavelet transform is applied to reduce the noise,followed by outlier detection based on the local outlier factor to extract the activity segment.Activity recognition is fulfilled by using the bi-directional long short-term memory that takes the sequential features into consideration.Experiments in through-the-wall environments achieve recognition accuracy>95%for six common activities,such as standing up,squatting down,walking,running,jumping,and falling,outperforming existing work in this field.

Outage performance of cognitive multisource multidestination relay networks with imperfect channel state information and interference from primary transmitter

Given imperfect channel state information(CSI)and considering the interference from the primary transmitter,an underlay cognitive multisource multidestination relay network is proposed.A closed-form exact outage probability and asymptotic outage probability are derived for the secondary system of the network.The results show that the outage probability is influenced by the source and destination number,the CSI imperfection as well as the interference from the primary transmitter,while the diversity order is independent of the CSI imperfection and the interference from the primary transmitter,yet it is equal to the minimum of the source and destination number.Moreover,extensive simulations are conducted with different system parameters to verify the theoretical analysis.

A joint optimization scheme of resource allocation in downlink NOMA with statistical channel state information

In a real communication scenario,it is very difficult to obtain the real-time channel state infor-mation(CSI)accurately,so the non-orthogonal multiple access(NOMA)system with statistical CSI has been researched.Aiming at the problem that the maximization of system sum rate cannot be solved directly,a step-by-step resource allocation optimization scheme based on machine learning is proposed.First,in order to achieve a trade-off between the system sum rate and user fairness,the system throughput formula is derived.Then,according to the combinatorial characteristics of the system throughput maximization problem,the original optimization problem is divided into two sub-problems,that are power allocation and user grouping.Finally,genetic algorithm is introduced to solve the sub-problem of power allocation,and hungarian algorithm is introduced to solve the sub-problem of user grouping.By comparing the ergodic data rate of NOMA users with statistical CSI and perfect CSI,the effectiveness of the statistical CSI sorting is verified.Compared with the orthogonal multiple access(OMA)scheme,the NOMA scheme with the fixed user grouping scheme and the random user grouping scheme,the system throughput performance of the proposed scheme is signifi-cantly improved.

基于CSI动态参数优化的人体轨迹追踪

为减少依靠单链路定位的时间、简化定位复杂度和优化参数选择,提出一种联合动态参数(joint dynamic parameter,JDP)算法。通过在单条链路采集信道状态信息(channel state information,CSI)数据,使用动静态信号分离技术减少数据的输入,简化参数的提取,联合多普勒频移加强动态参数特征,简化动态参数选择工作量,使用多个动态参数定位提高精准度。实验结果表明,相较现有的单链路定位技术,该方法提高了计算速度,计算时间显著减少,精准度有所提高。

大规模MIMO系统基于深度学习的轻量级CSI反馈方法

在大规模多输入多输出(multiple-input multiple-output,MIMO)系统中,基站需根据用户反馈的信道状态信息(channel state information,CSI)调制自适应编码提升谱效率。针对现有基于深度学习(deep learning,DL)的CSI反馈方法在用户端实际部署时可行性较差的问题,在卷积神经网络的基础上提出了一种轻量级的CSI反馈网络,并利用深度可分离卷积技术来减少反馈网络的参数量与计算量。考虑用户端实际部署,设计了在不同压缩比条件下以及在不同环境条件下的多任务融合反馈网络。仿真将所提方法与基于DL的ConvCsiNet和ShuffleNet反馈网络在归一化均方误差和参数量与计算量等方面进行对比与分析。结果表明,所提的反馈网络在保持较高CSI重构精度的前提下,可以极大减少用户端在实际部署时所需的参数量和计算量。

基于不完美CSI的水声自适应功率分配算法

水声信道面临带宽资源有限、环境复杂的问题,为提高水下通信速率,基于水声传感器网络的海洋应用提出自适应通信的需求。传统基于简单信噪比指标的自适应资源分配算法无法准确表述衰落信道的统计特征,利用强化学习和卷积神经网络预测信道的方法虽然可以提高一定信道状态信息(channel state information,CSI)的准确性,但这种方法需要长期的观测和大量的训练样本,不符合水声环境的实际情况。对比,构建了一种中继放大转发协作正交频分复用(orthogonal frequency division multiplexing,OFDM)通信的模型,解决了由于节点漂浮导致直接通信传输效率变低的问题,并提出一种在时延反馈CSI中基于OFDM的自适应功率比特分配算法,利用条件概率表征不完美的CSI,调整自适应通信参数,进行遍历容量最大化建模。仿真结果表明,该算法实现功率和比特的联合自适应分配,平均传输速率指标优于直接反馈CSI的功率分配算法,虽然略低于采用马尔可夫链预测信道的方法,但结合算法复杂度来看,所提算法更简单,更适合能量有限的水声传感器网络。