认知网络中一种基于频谱安全态势感知的路由方案

从构建准确的频谱态势图,实现频率资源充分利用的角度出发,针对目前在构建频谱态势图时没有考虑恶意用户(Malicious User,MU)存在的情况,结合克里金(Kriging)插值法估计空域内频谱干扰态势,通过地理位置检测方案识别MU,从而构建更准确、更安全的频谱态势图,并将其用于端到端的路由协议中。仿真结果表明,该方案能构建完整的频谱态势图,平均误差仅为0.106 dBm,能准确识别恶意用户,识别率高于80%;并且通过识别MU,在基于频谱态势图的路由过程中,可以减少路由跳数,增加可用频谱空间。     

基于CSS的认知无线电网络PUEA检测和缓解

针对认知无线电网络中的主用户仿真攻击问题,提出了一种基于能量检测的协作频谱感知方法.首先,假设存在一个智能攻击者,它能知道自己所处环境并可以选择不同的传输策略.然后,调整CSS规则和相关参数为能量探测器确定合适且有效的阈值.最后,利用Neyman-Pearson准则进行能量检测,从而检测出恶意攻击.仿真结果表明,相比传统的基于能量检测的频谱感知方法,该方法可以更好地减轻PUEA在频谱感知中的破坏作用.     

基于DRL的抗干扰电视频谱资源分配算法

将认知无线电与能量采集相结合,可以提高电视系统的频谱效率和能量效率。然而,由于无线信道的开放特性,频谱很容易受到恶意用户的干扰攻击,从而导致吞吐量下降。对此,将抗干扰频谱分配问题表述为没有任何先验知识的马尔可夫决策过程,然后提出一种基于深度确定性策略梯度(Deep Deterministic Policy Gradient,DDPG)的资源分配算法。在多种干扰环境下的仿真实验结果表明,该算法能够有效地减少恶意干扰带来的不利影响。     

基于可信度的半双工协同频谱检测

提出了基于可信度的半双工协同频谱检测方案.该方案利用贝叶斯方法来确定协同用户的可信度,分别就具有单协同用户和多协同用户的半双工协同频谱检测进行了研究,分析了可信度对协同频谱检测的影响.分析和仿真结果表明,本文策略不但能提高检测概率、降低检测时间,还能排除故障用户对检测性能的影响、减少恶意用户的干扰.     

智能感知辅助的快速动态频谱抗干扰方法

机器学习赋能的动态频谱抗干扰方法能够通过频谱感知学习干扰规律，自主优化抗干扰策略，适应动态复杂的频谱环境。然而，现有大部分研究假设环境中仅有恶意干扰而不考虑其他通信系统的存在，且所提算法复杂度和算力要求较高。针对以上问题，考虑主用户和恶意干扰同时存在且用频规律未知的场景，设计了基于并行学习的智能感知算法和基于预先学习的动态频谱接入算法，无须通过随机探索频谱环境即可学习可用信道变化规律。仿真结果表明，所提算法在动态干扰攻击下能在线快速找到最优信道选择策略，在完美避开动态干扰的同时不对主用户造成无意干扰。     

无线认知传感器网络协同频谱感知安全模型和攻击模型研究

无线认知传感器网络(WCSN)中的协同频谱感知技术能够大大提高检测主用户传输的概率.但是,在当前的协同频语感知方案中,通常假定所有协同次用户都是诚信可靠的,因而恶意次用户可能会通过报告错误检测结果来攻击网络.本文分析了无线认知传感器网络(WCSN )的基本概念和优势,给出了无线认知传感器网络体系结构,提出了协同频谱感知...     

认知无线电多时隙联合频谱感知方法及优化

为了降低认知无线电中次用户对主用户的干扰,提出了次用户通过多个时隙的能量感知联合检测主用户的频谱感知方法。每个传输帧被分成若干个时隙,次用户在每个时隙的开始进行能量感知,通过合并多个时隙的感知结果,提高次用户对主用户的检测性能。为了最大化次用户的频谱效率,将主用户状态建模为二维马尔科夫随机过程,并根据该过程优化单时隙频谱感知时间。仿真结果表明：相比Liang的“先听后传”频谱感知方法,提出方法仅牺牲14％的频谱效率,却使干扰概率降低了28％;随着信噪比增加,提出方法的最大频谱效率逐渐接近“先听后传”频谱感知方法。     

一种SSDF攻击检测的证据理论合作频谱感知算法

现有的基于证据理论的合作频谱感知认为所有认知用户都是诚实的,没有考虑恶意用户的存在。当恶意用户篡改本地感知的结果,发送错误的数据到数据融合中心,将会降低合作频谱感知的性能,这种攻击称为频谱感知数据篡改（spectrum sensing data falsification,SSDF）攻击。由于恶意用户发送的证据与其它认知用户的证据存在差别,本文使用Jousselme距离来衡量证据的可信度,提出一种SSDF攻击检测算法。数据融合中心接收所有认知用户的证据,让可信度低的证据不参与融合判决,可信度高的证据根据可信度进行加权融合。仿真结果表明,所提出的SSDF攻击检测算法在恶意用户发起SSDF攻击时可以很好的改善频谱感知性能。      

无线认知传感器网络特征模板匹配算法研究

频谱感知是实现认知无线电应用、构建WCSN(无线认知传感器网)的核心技术,也是保护主用户免受有害干扰、提高WCSN自身频谱资源利用率的重要前提。当前的频谱感知技术主要有匹配滤波器检测、特征模板匹配和能量检测。文章给出WCSN频谱感知的假设检验,分析主用户信号主特征向量的求解方法,描述PCA(主成分分析)框架下检测算法的实现流程。频谱感知实验结果表明,PCA方法比对应线性方法性能优越。     

协作频谱感知中的SSDF攻击及其对策研究

协作频谱感知是认知无线电中实现频谱有效利用的关键环节,但协作频谱感知由于其聚集感知数据的特点,为恶意用户提供了可乘之机。恶意用户可以通过实施频谱感知数据伪造(spectrum sensing data falsification,SSDF)攻击的方式,使认知无线电系统不能如实地根据外界环境进行动态的频谱切换。为深入理解SSDF攻击,掌握其防御对策的研究现状及发展趋势,从攻击手段、攻击形式和攻击策略3个角度细化出具体的SSDF攻击类型,总结了目前SSDF攻击的典型防御对策。同时,针对目前相关研究中所存在的问题,明确了一些有待继续研究的方向。     

Towards trustworthy collaboration in spectrum sensing for ad hoc cognitive radio networks

Cognitive radio networks (CRN) make use of dynamic spectrum access to communicate opportunistically in frequency bands otherwise licensed to incumbent primary users such as TV broadcast. To prevent interference to primary users it is vital for secondary users in CRNs to conduct accurate spectrum sensing, which is especially challenging when the transmission range of primary users is shorter compared to the size of the CRN. This task becomes even more challenging in the presence of malicious secondary users that launch spectrum sensing data falsification (SSDF) attacks by providing false spectrum reports. Existing solutions to detect such malicious behaviors cannot be utilized in scenarios where the transmission range of primary users is limited within a small sub-region of the CRN. In this paper, we present a framework for trustworthy collaboration in spectrum sensing for ad hoc CRNs. This framework incorporates a semi-supervised spatio-spectral anomaly/outlier detection system and a reputation system, both designed to detect byzantine attacks in the form of SSDF from malicious nodes within the CRN. The framework guarantees protection of incumbent primary users’ communication rights while at the same time making optimal use of the spectrum when it is not used by primary users. Simulation carried out under typical network conditions and attack scenarios shows that our proposed framework can achieve spectrum decision accuracy up to 99.3 % and detect malicious nodes up to 98 % of the time.     

Fast and Accurate Detection of Malicious Users in Cooperative Spectrum Sensing Network

Cooperative spectrum sensing allows users to detect available spectrum and utilize it. However, it is known that a few users can easily affect the cooperative decision at the fusion center by reporting falsified sensing data. The problem of detecting malicious users in cooperative spectrum sensing has been studied by numerous researchers. The basic approach is to compute the credibility of the reported data of each user and declare those with low credibility (below some threshold) as malicious. The computation of credibility can be based on several things such as pattern of historic behavior, entropy of the reported data, signal-to-noise ratio, etc. The credibility of the reported data of some user node can be expressed as an attack probability of that node. A node with higher attack probability will have lower credibility while a node with lower attack probability a higher credibility. Then, the problem of computing credibility of a node becomes the problem of computing attack probability of a node given a set of reported data. We can enumerate a list of all possible attack probability vectors for a set of nodes and the list of all possible energy level vectors of the channel for the observed time period, and compute which combination of attack probability and channel energy level vector can have the maximum probability to produce the reported channel energy level. However, the search space is quite large and grows exponentially as the number of user nodes and the number of time slots to observe increase. In this paper, we suggest algorithms that reduce the search space considerably and detect malicious users in linear time instead of exponential time. The suggested algorithms have been implemented and show promising results.

     

Throughput of cognitive radio networks with improved energy detector under security threats

Spectrum sensing in cognitive radio networks (CRNs) is subjected to some security threats such as primary user emulation (PUE) attack and spectrum sensing data falsification (SSDF) attack. In PUE attack, a malicious user (MU_PUE) transmits an emulated primary signal throughout the spectrum sensing interval to secondary users (SUs) to forestall them from accessing the primary user (PU) spectrum bands. In SSDF attack, malicious users (MU_SSDF) intentionally report false sensing decisions to the fusion center (FC) to influence the overall decision. While most of the existing literatures have studied the effects of these 2 types of attacks separately, the present paper evaluates the secondary network performance in terms of throughput under both the PUE and SSDF attacks with improved energy detectors (IEDs) where SU's spectrum access is hybrid, ie, either in overlay or in underlay mode. An analytical expression on throughput of SU under the simultaneous influence of both of these attacks is developed. Impact of several parameters such as IED parameter, attacker probabilities, and attacker strength on the throughput of SU is investigated. Performance of the present scheme is also compared with only PUE and only SSDF attacks. A simulation test bed is developed in MATLAB to validate our analytical results.     

Fault-Tolerant Cooperative Spectrum Sensing Scheme for Cognitive Radio Networks

In order to solve the uncertainty resulting from shadowing effect and resist the attack from malicious cognitive radio (CR) users, we propose a fault-tolerant cooperative spectrum sensing scheme for CR networks, where an energy detection-based local spectrum sensing is performed at each CR user, a coefficient is used to weight each CR user’s sensing result, a linear weighted fusion process is performed at the fusion center (FC) to combine received sensing results. For a fault-tolerant cooperative spectrum sensing scheme, the most important issue is to distinguish whether the CR user is reliable or not. In this paper, a reputation-based cooperative mechanism is presented to alleviate the influence of the unreliable sensing results from CR users suffering shadowing and the false sensing data from malicious CR users on the detection result at the FC. In proposed fault-tolerant cooperative scheme, each cooperative CR user has a reputation degree which is initialized and adjusted by the FC and used to weight the sensing result from the corresponding user in the fusion process at the FC. And then, two reputation degree adjusting methods are presented to manage the reputation degree of each CR user. Simulation results show that the proposed scheme can not only weaken the harmful influence caused by malicious CR users, but also alleviate the corrupted detection problem resulting from destructive channel condition between the primary transmitter and the CR user. Moreover, the detection performance of the fault-tolerant cooperative scheme, which has a feasible computational complexity and needs no instantaneous SNRs, is close to that of the optimal scheme.     

Weight sequential log-likelihood ratio detect algorithm with malicious users removing

Due to the openness of the cognitive radio network, spectrum sensing data falsification (SSDF) can attack the spectrum sensing easily, while there is no effective algorithm proposed in current research work, so this paper introduces the malicious users removing to the weight sequential probability radio test (WSPRT). The terminals' weight is weighted by the accuracy of their spectrum sensing information, which can also be used to detect the malicious user. If one terminal owns a low weight, it can be treated as malicious user, and should be removed from the aggregation center. Simulation results show that the improved WSPRT can achieve higher performance compared with the other two conventional sequential detection methods under different number of malicious users.     

Recent advances and future challenges of four key resources cooperation in cognitive radio network

Cognitive radio (CR) can improve spectrum utilization by spectrum sharing or cooperation between the primary user and secondary users.It is well known that energy,interference and relay are also three key resources in cognitive radio network (CRN).Energy cooperation or sharing between the primary user and secondary user will further promote energy efficiency.Energy harvesting from RF interference signal can turn bane (interference) into a boon (green energy).Secondary user relay data of the primary user can enhance QoS of the primary user,also get some opportunities for their own transmission.Thus,four resources cooperation (spectrum,energy,interference,relay) in CRN will improve simultaneously both spectrum efficiency and energy efficiency,and also increase throughput and QoS.The overviews for collaborative utilization problems of four key resources in CRN was given.Firstly,collaboration models of four key resources were analyzed.Then,recent research advances were summarized,including three kinds of resources collaborative utilization (both energy harvesting and relay transmission,both energy cooperation and energy harvesting) and four resources collaborative utilization (simultaneous relay transmission,energy harvesting and energy cooperation) in CRN.Further,some potential challenges of four key resources cooperation in CRN were discussed.Finally,some key future research directions was concluded.     

可信的反馈声誉保护协作频谱感知软判决算法

协作频谱感知中信任机制的引入,起到了抑制恶意用户频谱感知数据伪造( SSDF)攻击行为的作用。然而,数据融合中心不加区分地接收协作感知结束后的反馈信息,为恶意用户带来了实施“掺沙子”攻击的机会。恶意用户向数据融合中心反馈错误的主用户频谱状态,使信任机制不能得出准确的信任值。为此,提出了一种基于反馈声誉的信任机制,考虑反馈中的个体性特征,引入反馈声誉的思想来量化认知用户信任值。同时,将信任值量化结果用于权重经典软判决算法———序贯概率比检测( SPRT)算法,消除SSDF恶意用户参与软判决数据融合的影响,形成可信序贯概率比检测算法( FSPRT)。仿真结果表明FSPRT算法的性能优于传统SPRT算法,能有效降低网络信任值计算误差,并保持较好的感知性能。     

基于合约制的认知网络协作频谱共享模型

为了提高认知无线电网络中主用户的能效和谱效,提出了一种基于合约制的协作频谱共享新模型。在该模型中,当主用户信道质量差时,次用户为其提供中继服务;作为回报,在主用户信道质量好时,次用户可以在一定的干扰约束下以underlay的模式和主用户共享频谱。在此框架下引入经济学中的合约理论,将主次用户间的协作问题建模为合约设计问题,并建立了主次用户评价合约收益的效用函数,将提高主用户的能效和谱效问题转变为设计合约使主用户获得最大效用的优化问题,并利用差分进化算法对该问题求解。最后,在不同的环境下,将主用户协作与不协作获得的效用进行对比,结果表明,主用户在没有额外投入频谱资源的条件下,不仅节省了发送功率,而且有效地提高了数据传输速率,进而提高了能效和谱效。     

Ad hoc认知无线网络的主用户感知定位方案

为了对频谱资源进行合理的运用,在认知无线Ad hoc网络中,各次用户必须准确地对主用户信号进行感知,以防止恶意攻击者发起的模仿授权用户攻击。本文基于对信号的波达方向(DOA)估计,提出了一种存在系统误差下对主用户的感知定位算法,仿真证明,该方案能准确地对主用户信号进行定位;当信道环境恶劣、信噪比较低时,本文通过引入环境系数,提出了一种灵活性较好的协作加权检测策略。该策略中每个次用户都可以做出对主用户信号的最终判决,解决了以往协作感知方案中由于权力过分集中而引起的潜在安全性问题。