基于GRU和LSTM组合模型的车联网信道分配方法

针对车联网中高通信需求和高移动性造成的车对车链路(Vehicle to Vehicle, V2V)间的信道冲突及网络效用低下的问题,提出了一种基于并联门控循环单元(Gated Recurrent Unit, GRU)和长短期记忆网络(Long Short-Term Memory, LSTM)的组合模型的车联网信道分配算法。算法以降低V2V链路信道碰撞率和空闲率为目标,将信道分配问题建模为分布式深度强化学习问题,使每条V2V链路作为单个智能体,并通过最大化每回合平均奖励的方式进行集中训练、分布式执行。在训练过程中借助GRU训练周期短和LSTM拟合精度高的组合优势去拟合深度双重Q学习中Q函数,使V2V链路能快速地学习优化信道分配策略,合理地复用车对基础设施(Vehicle to Infrastructure, V2I)链路的信道资源,实现网络效用最大化。仿真结果表明,与单纯使用GRU或者LSTM网络模型的分配算法相比,该算法在收敛速度方面加快了5个训练回合,V2V链路间的信道碰撞率和空闲率降低了约27%,平均成功率提升了约10%。     

基于深度学习的信道估计技术研究进展

信道估计是接收机基带信号处理的关键,直接决定了无线通信系统的通信服务质量。传统的信道估计方法已经不能满足日益复杂和个性化的现代通信需求,同时人工智能技术特别是深度学习已被应用于无线通信物理层并带来了良好的通信性能增益。为系统地总结上述研究成果,并探讨未来的技术发展趋势,从数据驱动和模型驱动两方面分别对基于深度学习的信道估计方法进行了分析和归纳,并且描述了其中代表性算法,最后探讨了基于深度学习的信道估计的研究挑战与趋势。     

Intelligent channel estimation in millimeter wave massive MIMO communication system using hybrid deep learning with heuristic improvement

The major goal of millimeter wave (mmWave) multiple-input multiple-output (MIMO) systems is to get effective channel state information (CSI). Most of the recent works use nuclear norm theory for recovering the low-rank scheme of channels. Some suboptimal solutions to the rank minimization problem can occur while addressing the nuclear norm-based convex problem, which degrades the accuracy of channel estimation. Some works recover the channel with the assumption of the mmWave channel using an over-complete dictionary. On the other hand, the accuracy of available CSI may openly influence the efficiency of mmWave communications. The main intention of this paper is to develop an enhanced channel estimation model with an optimized hybrid deep learning model. Here, the integration of deep neural network (DNN) and long short-term memory (LSTM) form the hybrid deep learning model termed optimized D-LSTM, which is modified by the opposition searched exploration-based Harris hawks optimization (OE-HHO). The input to the proposed hybrid deep learning is taken as the correlation among the received signal vectors and the measurement matrix for predicting the beam space channel amplitude. Finally, the successful channel estimation is observed by deep hybrid learning by the experimental outcomes, which also demonstrate that the proposed channel estimation model overwhelms the conventional models in terms of Normalized Mean-Squared Error (NMSE) and spectral efficiency. The experimental results show that the designed OE-HHO method obtains 9.2%, 8.9%, 8.65%, and 0.47% progressed than DA, DHOA, GWO, and HHO, respectively. Therefore, higher efficiency is observed by OE-HHO based mmWave MIMO communication system.     

Optimal insertion of pilot symbols for transmissions over time-varying flat fading channels

Two major training techniques for wireless channels are time-division multiplexed (TDM) training and superimposed training. For the TDM schemes with regular periodic placements (RPPs), the closed-form expression for the steady-state minimum mean square error (MMSE) of the channel estimate is obtained as a function of placement for Gauss-Markov flat fading channels. We then show that among all periodic placements, the single pilot RPP scheme (RPP-1) minimizes the maximum steady-state channel MMSE. For binary phase-shift keying (BPSK) and quadrature phase-shift keying (QPSK) signaling, we further show that the optimal placement that minimizes the maximum uncoded bit error rate (BER) is also RPP-1. We next compare the MMSE and BER performance under the superimposed training scheme with those under the optimal TDM scheme. It is shown that while the RPP-1 scheme performs better at high SNR and for slowly varying channels, the superimposed scheme outperforms RPP-1 in the other regimes. This demonstrates the potential for using superimposed training in relatively fast time-varying environments.     

结合反卷积和扩张卷积的信道估计算法

信道估计作为无线通信的关键,近年来成为相关领域的研究热点。本文针对正交频分复用（Orthogonal Frequency Division Multiplexing,OFDM）系统下传统信道估计算法性能难以满足复杂场景的通信需求、受噪声影响大等问题,提出了一种基于反卷积网络及扩张卷积网络信道估计的深度学习方法。该方法利用信道的相关性构建了一个轻量级的反卷积网络,利用少数几层反卷积操作来逐步实现信道插值与估计,在较低的复杂度下较好地实现了信道估计。为改善估计性能,进一步构建了一个扩张卷积网络来抑制信道噪声,提高信道估计的准确度。仿真结果表明,在不同信噪比条件下,本文提出的基于反卷积及扩张卷积的深度学习方法比传统方法具有更低的估计误差,且复杂度较低。      

一种面向LTE-V2X的联邦学习信道估计算法

基于深度学习的信道估计方法中,训练网络模型需要大量的数据运算,且所有用户数据都需要集中上传至服务器上,存在隐私泄漏的隐患.针对上述问题,提出了一种基于联邦学习的LTE-V2X(Long Term Evolution-Vehicle to Everything)信道估计算法,采用CNN-LSTM-DNN(Convolutional Neural Network-Long Short Term Memory-Deep Neural Network)模型对时变的信道进行估计,并将学习网络模型所需要的计算分配到车载用户中,在降低道旁基站负载的同时也保护了车载用户数据的隐私.仿真结果表明,基于联邦学习的信道估计算法在车载用户高速移动的场景下,较传统的信道估计算法平均有10 dB以上的归一化均方误差(Normalized Mean Square Error,NMSE)增益以及3 dB以上的误码率(Bit Error Rate,BER)增益,且较集中式学习算法相比,NMSE性能差距在3 dB以内;BER性能差距在1 dB以内,所提算法能够有效追踪时变的信道,且与集中式学习算法相比仅损失了极少的性能.     

60GHz室内天线位置优化与天线选择策略

采用多天线技术的60 GHz无线通信被认为是未来室内场景下高数据率宽带无线接入一种有前途的解决方案。参考802.11ad工作小组采用的60 GHz信道模型,根据60 GHz路径损耗公式提出的分布式天线系统间距的优化方案可以保证房间内功率覆盖的均匀性。在此基础上,进一步分析60 GHz多天线系统在不同链路条件下的信道条件数关系,提出一种基于信道条件数的天线选择策略。该策略利用信道条件数判断信道性能,既保证了通信质量,又适时降低系统总功率,切换简单、易于操作,适用于60 GHz室内无线通信。     

应用深度学习的信号解调

针对噪声干扰信道下的信号解调问题,提出了应用深度学习的信号识别方法,通过识别信号完成信号解调.深层置信网络使用受限波尔兹曼机为基本单元,设计针对通信信号识别的多层深层置信网络.通信信号首先变换为特定表征序列,并以此构建完备的训练集合对深度置信网络进行逐层的无监督学习和全局有监督的微调反馈学习,在深层置信网络的权重参数优化过程中实现对通信信号的特征提取与识别.仿真实验表明,与传统调制信号解调方法相比,应用深度学习的信号解调方法的检测性能有约0.4 dB的提升.     

基于深度学习的单载波频域均衡算法研究

单载波频域均衡(Single-Carrier Frequency-Domain Equalization,SC-FDE)是一种有效的抗码间干扰的算法,在无线通信系统中得到了广泛的应用。传统线性SC-FDE算法主要包括信道估计、噪声功率估计和信道均衡三个模块,其中每个模块都是单独优化的。为了联合优化这三个模块,本文提出了一种基于深度学习的SC-FDE算法。为了减少网络收敛所需的训练数据量,本文为SC-FDE中的三个模块分别设计了一个子网络。此外,本文还提出了一种训练机制,通过平等地对待每条无线路径,提高了所提算法的信道泛化能力。仿真结果表明,所提算法可以在较小的训练数据集下收敛,且具有鲁棒的信道泛化能力,与基于最小二乘信道估计和最小均方误差信道均衡的SC-FDE算法相比,所提算法具有更优的误码率性能。      

MISO-CDMA多径系统中空时分组码的译码方法

空时编码是实现宽带无线数据通信和下一代移动通信系统的一种极有潜力的技术。为有效的将空时分组码应用到多径衰落环境下的码分多址系统，以充分利用多个路径的信号能量，现提出了一种多径环境下空时分组译码的新方法。由于空时分组码译码与信道估计紧密相关，为此本文对多径信道估计以及信道估计误差对本方法产生的影响作研究。仿真结果表明，采用多路径译码方法可以明显提高系统的误码性能。     

Deep learning-based channel estimation for wireless ultraviolet MIMO communication systems

To solve the problems of pulse broadening and channel fading caused by atmospheric scattering and turbulence, multiple-input multiple-output(MIMO) technology is a valid way. A wireless ultraviolet(UV) MIMO channel estimation approach based on deep learning is provided in this paper. The deep learning is used to convert the channel estimation into the image processing. By combining convolutional neural network(CNN) and attention mechanism(AM), the learning model is designed to extract the depth f...     

Hybrid Global Sensitivity Analysis Based Optimal Attribute Selection Using Classification Techniques by Machine Learning Algorithm

In this study, we focus on realizing channel estimation using a fully connected deep neural network. The data aided estimation approach is employed. We assume the transmission channel is Rayleigh and it is constant over the duration of a symbol plus pilot transmission. We develop and tune the deep learning model for various size of pilot data that is known to the receiver and used for channel estimation. The deep learning models are trained on the Rayleigh channel. The performance of the model is discussed for various size of pilot by providing Bit Error Rate of the model. The Bit Error Rate performance of the model is compared to theoretical upper bound which shows that the model successfully estimates the channel.

     

跳频/多载波频率分集/扩频多址无线通信系统的研究

本文提出了一种跳频／多载波频率分集／扩频多址（ＦＨ／ＭＣＦＤ／ＳＳＭＡ）无线通信系统,给出了ＦＨ／ＭＣＦＤ／ＳＳＭＡ系统的发送和接收模型,对判惟变量统计特性进行了分析,然后对峰窝系统反向链路在理想定时和信道估计条件下用户平均接收误码率进行了仿真。结果表明,ＦＨ／ＭＣＦＤ／ＳＳＭＡ蜂窝通信系统具有较好的抗多径衰落能力,同单载波ＦＨ／ＳＳＭＡ系统相比其误码性能和频谱效率有显著改善。     

基于1D-Concatenate的信道估计DNN模型优化方法

为提高DNN模型在无线通信中信道估计精度,提出一种基于1D-Concatenate的信道估计DNN模型优化方法。该方法将Concatenate进行一维（1D）数据转换,以跳跃连接的方式引入DNN模型,抑制梯度消失问题,运用1D-Concatenate恢复网络训练过程中丢失的数据特征,提高DNN信道估计精度。为验证优化方法的有效性,选取较典型的基于DNN的无线通信信道估计模型进行对比仿真实验。实验结果表明,本文提出的优化方法对已有DNN模型的估计增益提升可达77.10%,在高信噪比下信道增益提升可达3 dB。该优化方法能有效提高DNN模型在无线通信中的信道估计精度,特别是高信噪比下提升效果显著。     

深度学习光流计算技术研究进展

图像序列光流计算是图像处理与计算机视觉等领域的重要研究方向.随着深度学习技术的快速发展,以卷积神经网络为代表的深度学习理论与方法成为光流计算技术研究的热点.本文主要对深度学习光流计算技术研究进行综述,首先介绍了有监督学习、无监督学习和半监督学习的光流计算网络模型与训练策略,然后重点阐述并分析了不同网络模型优化方法.针对光流计算模型的评估问题,分别介绍了Middlebury、MPI-Sintel和KITTI等数据库及评价基准,并对不同类型深度学习和传统变分光流模型进行对比与分析.最后,总结了深度学习光流计算技术在模型复杂度与泛化性、光流估计鲁棒性、小样本训练准确性等方面的关键技术问题,并指出了可能的解决方案与研究思路.     

基于深度强化学习的无人机间通信链路智能决策

由于无人机组网灵活、快速、低成本的特性,空中基站被视为在未来无线通信中有前景的技术。无人机集群可以通过相互协调和合作,完成的复杂任务,具有重大的研究和实用价值,而无人机间的高效通信是当下面临的重大挑战。为了在满足无人机间通信速率的前提下,尽可能节省发射功率,本文提出基于深度强化学习的集群方案和功率控制的智能决策算法。首先,本文设计了三种无人机集群方案,以对地面用户提供无缝的无线覆盖;然后,本文提出了基于深度Q网络（Deep Q-network）算法的集群方案和功率控制决策算法,用深度神经网络输出不同条件下联合决策的无人机集群方案和发射功率,并研究了重要性采样技术,提高训练效率。仿真结果表明,本文提出的深度强化学习算法能够正确决策无人机集群方案和发射功率,与不带强化学习的深度学习（Deep Learning Without Reinforcement Learning, DL-WO-RL）算法相比,用更低的发射功率满足无人机之间的通信速率要求,并且重要性采样技术能够缩短DQN算法的收敛时间。      

面向RIS-mmWave系统的机器学习辅助高效波束训练及信道估计方法

面向B5G及6G无线通信系统的高速无线信息传输,本文研究了智能超表面辅助毫米波（RIS-mmWave）系统的高效能波束训练及信道估计方法。特别地,基于RIS-mmWave波束管理及有效信道获取的内生关联性,本文创新性地提出一种机器学习辅助的RIS-mmWave系统高效波束训练及信道估计方法。具体而言,在第一阶段,设计了一种新颖的半监督学习模型,实现位置信息辅助的在线快速波束训练,并且免估计地直接获取粗略角度域信息以驱动精细化信道估计;在第二阶段中,提出半监督学习辅助的压缩感知级联信道估计算法,利用半监督学习模型直接输出的粗略角度域信息驱动块正交匹配追踪算法进行信道估计。仿真结果表明,所提波束训练及信道估计方法在系统开销和信道估计误差等方面的性能均优于代表性参考方案。     

Deep Learning Based Physical Layer Security of D2D Underlay Cellular Network

In order to improve the physical layer security of the device-to-device(D2D)cellular network,we propose a collaborative scheme for the transmit antenna selection and the optimal D2D pair establishment based on deep learning.Due to the mobility of users,using the current channel state information to select a transmit antenna or establish a D2D pair for the next time slot cannot ensure secure communication.Therefore,in this paper,we utilize the Echo State Network(ESN)to select the transmit antenna and the Long Short-Term Memory(LSTM)to establish the D2D pair.The simulation results show that the LSTMbased and ESN-based collaboration scheme can effectively improve the security capacity of the cellular network with D2D and increase the life of the base station.     

HPLC&RF双模系统中无线通信的导频设计

随着高速电力线载波通信(High-speed Power Line carrier Communication,HPLC)技术在电力物联网中的推广和应用,其面临的组网孤岛/孤点、停电事件上报成功率低、单跳通信距离短等问题逐步显现.无线(Radio Frequency,RF)通信能够有效地解决这些问题,从而成为了HPLC的有力补充.对于HP LC&RF双模系统中的无线通信,现有的导频设计没有充分考虑其典型应用信道的时频相关性,为此提出了一种新的导频设计方案.该方案增加了导频的频域密度,以更好地适应频率选择性高的信道;同时基于所有典型信道随时间变化缓慢的特性,降低了导频的时域密度;另外,重新设计了导频的时频位置,以进一步降低信道估计的复杂度.新方案的导频开销为原方案的1/2.仿真结果表明,所提方案的性能均优于原方案,且适用的信道估计方法简单,计算复杂度低,易于实现,能够更好地满足双模系统的推广应用需求.     

Performance analysis for relay networks with hierarchical support vector machines

In this paper, we consider a cooperative relay scheme for a mobile network with MIMO technology. The channel capacity for two well‐known relaying schemes are investigated: analogue relaying (amplify and forward) and digital relaying (decode and forward) from a mobile device to the base station through a relay node. In order to further increase the channel capacity, we propose an efficient hierarchical procedure based on support vector machine, namely hierarchical support vector machines (HSVM), to estimate the wireless networks condition approximately and design two ways (matched filter and minimum mean square error filter) of increasing the channel capacity according to the estimated wireless network condition. The proposed HSVM can estimate the wireless networks condition in much shorter time compared with the traditional minimum mean square error scheme without incurring much estimation error, which is spatial, useful for delay sensitive communication. For digital relaying, the effect of imperfect channel decode is also addressed. Our numerical results demonstrate the reduction of estimation complexity by adopting HSVM and the significant improvement of network capacity by applying the matched filter weight at relay nodes according to the network estimation. Copyright © 2011 John Wiley & Sons, Ltd.