稳健精细抗差异性频谱感知技术研究

发布时间：2018-07-28 14:37

【摘要】：频谱感知技术是研究和发展认知动态系统的关键基础要素,然而随着无线通信、电子侦察以及电子对抗等技术的迅速发展,电子设备的种类层出不穷,各种无线电信号功率电平差异越来越大,通信信号体制、探测方式以及干扰模式也越来越多样,使得空间电磁环境越来越复杂,这给频谱感知技术带来许多亟待解决和必须持续关注的新需求:1.稳健频谱感知在实用的认知动态系统中,频谱感知技术必须具备在极低信噪比和无线信道严重衰落电磁背景下,进行实时盲频谱感知的能力,这给经典的频谱感知算法感知的稳健性带来巨大的挑战。2.精细频谱感知新一代无线通信、监测、侦察、对抗系统都或多或少地朝着超宽带、短时突发以及多交互目标的趋势进一步推进,通信电子行业的这一技术发展趋势需要认知动态系统具备超宽带多目标实时感知的能力,这为频谱感知技术引入实时地在超宽带内精细化分析的技术难题。3.深度频谱感知针对某些认知动态系统,如认知雷达、认知电子对抗、认知无线电等,需要具备对无线电信号深度分析的能力,以获取任意目标信号的射频特性、通信和干扰制式、调制类型、波形成形、来波方向以及位置等等众多关键信息,从而实现系统的联合优化设计。然而,目前国内外频谱感知的技术和设备,极少具有深度感知的能力,这也将是频谱感知技术需要长期演化进步的一个方向。4.抗差异性协作频谱感知背景噪声是由地面噪声、大气噪声、降雨噪声、人为噪声、干扰噪声以及信号检测接收机热噪声等多种噪声形式复合而成,背景噪声电平往往在时间、地点以及频率等多维域上呈现出异常复杂的高动态起伏变化的特点。为了使协作频谱感知的结果更为准确和可靠,协作频谱感知技术必须具备适应参与协作节点背景噪声电平高动态变化的能力,来对抗和削弱各节点自身差异性带给最终协作结果的不良影响。本文针对认知动态系统中频谱感知技术面临的新需求展开研究,论文主要研究成果如下:1.频谱感知的系统模型常被简单地塑造成二元假设的问题,往往忽略了衰落信道系数、信号码元速率以及白噪声带宽之间的相互影响,本文构建的系统模型,充分分析了三者之间的关联,并重点考虑了平坦慢衰落信道引入的样本之间的相关性,保证了系统模型的精确性。2.针对现存频谱感知算法面临的稳健性需求,本文从频域信号处理的角度,利用归一化纯量变换的原理,设计出基于归一化谱的信号检测算法,该算法依据傅里叶变换的渐进正态性和相互独立性计算功率谱统计特性,利用监测频带内部分谱线强度和与全部谱线强度和的比值作为检验统计量,该算法的判决门限只与频谱感知算法的参数配置有关,而与节点的噪声方差无关,可有效克服噪声不确定度对频谱感知性能的影响,固定信噪比,算法的频谱感知性能不受噪声电平改变的影响,应用于高斯白噪声和平坦慢衰落信道中,可在较宽的信噪比范围内获得较优越的频谱感知性能。3.针对频谱感知精细化的需求,现有常规超外差窄带频谱感知技术无法快速精确完成超宽带内多目标的频谱感知,本文进行面向离散频段的实时多目标并行频谱感知研究,利用多通道-多相化的结构进行超宽频段功率谱的计算,然后并行计算出各离散频段的归一化功率谱,通过循环执行正向和反向搜索,检测出通信带宽内被占用的多个频隙,在一个搜索循环中,首先执行正向判决,检测出瞬时功率参差不齐、差异较大的子带信号,然后执行反向判决,检测出正向判决中漏检的类似梳状信号类型的子带信号,完成对带内的多个目标的并行频谱感知。4.针对参与协作节点自身参数的差异性,论文首先设计了一种更具有普适性的协作频谱感知算法,该算法将各节点本地归一化谱上传融合中心,克服噪声电平动态变化在时域、空域和频域上的效应积累,然后在融合中心采用等增益平均或者最优化加权平均的方式计算检验统计量,可有效消除背景噪声高动态变化对信号频谱感知性能的影响。5.设计和硬件实现了频谱传感器节点,然后在完成频谱传感器组网的基础上,设计了一种高精度同步数据采集方法,最后对比了归一化谱频谱感知算法和能量频谱感知算法的单点和协作感知实测性能。
[Abstract]:Spectrum sensing technology is the key basic element for the research and development of cognitive dynamic systems. However, with the rapid development of wireless communications, electronic reconnaissance and electronic countermeasures, the types of electronic devices emerge in endlessly, the power levels of various radio signals are more and more different, the more the communication signal system, the detection mode and the interference mode are also. The more and more diversity, the more and more complex space electromagnetic environment, which brings a lot of new needs to be solved and must continue to pay attention to the spectrum sensing technology: 1. robust spectrum sensing in the practical cognitive dynamic system, spectrum sensing technology must have the real-time blind frequency under the extremely low signal to noise ratio and the wireless channel severe fading electromagnetic background. The ability of spectral perception brings great challenge to the robustness of classical spectrum sensing algorithms..2. fine spectrum sensing new generation of wireless communication, monitoring, reconnaissance, and antagonism are more or less moving towards ultra wideband, short-time burst and multi interactive target, and this technology development trend of communication electronics industry is more or less. Cognitive dynamic systems have the ability to have ultra wideband multi-target real-time perception. This is a technical problem for spectral sensing to introduce real time fine analysis in ultra wideband..3. deep spectrum sensing, such as cognitive radar, cognitive electronic countermeasures, cognitive radio, and so on, needs to have the depth of radio signals. The ability to analyze the radio frequency characteristics of any target signal, communication and interference mode, modulation type, waveform forming, wave direction and position and so on, so as to achieve the joint optimization design of the system. However, at present, the technology and equipment of spectrum sensing at home and abroad are very few with the ability of depth perception, which will also be the spectrum. The perceptual technology needs a direction of long-term evolution and progress..4. anti heterosexual cooperative spectrum sensing background noise is composed of ground noise, atmospheric noise, rain noise, artificial noise, interference noise, and the thermal noise of signal detection receiver. The background noise level is often multidimensional in time, location and frequency. In order to make the results of cooperative spectrum sensing more accurate and reliable, the cooperative spectrum sensing technology must have the ability to adapt to the high dynamic changes of the background noise level of the participating cooperation nodes, to counter and weaken the bad results of the nodes themselves to the final cooperation results. The main research results of this paper are as follows: 1. the system model of spectrum sensing is often simply molded into the two element hypothesis, often neglecting the fading channel coefficient, the signal code rate and the interaction between the white noise bandwidth, and the paper. The built system model fully analyzes the correlation between the three, and focuses on the correlation between the samples introduced by the flat slow fading channel, which ensures the accuracy of the system model to meet the robustness requirements of the existing spectrum sensing algorithm. This paper uses the principle of normalized pure transformation from the angle of frequency domain signal processing. A signal detection algorithm based on normalized spectrum is taken into account. According to the asymptotic normality and mutual independence of Fu Liye transform, the algorithm is used to calculate the statistical characteristics of power spectrum, using the intensity of spectral lines in the monitoring band and the ratio of all spectral lines to the intensity of the spectrum. The threshold of the algorithm is only matched with the parameters of the spectrum sensing algorithm. It has nothing to do with the noise variance of the node. It can effectively overcome the influence of the noise uncertainty on the spectrum sensing performance. The fixed signal to noise ratio, the spectrum sensing performance of the algorithm is not affected by the change of noise level. It can be applied to Gauss white noise and flat slow fading channels, and a better spectrum sense can be obtained in a wider range of signal to noise ratio. The existing conventional ultra heterodyne narrowband spectrum sensing technology can not quickly and accurately complete the spectrum sensing of the UWB multiple targets. In this paper, the real-time multi-target parallel spectrum sensing research oriented to discrete frequency band is studied and the multi channel multiphase structure is used to calculate the power spectrum of the ultra wide band.3.. Then the normalized power spectrum of each discrete frequency band is calculated in parallel, and the multiple frequency gaps are detected in the communication bandwidth by cyclic forward and reverse search. In a search cycle, the forward decision is performed first, and the instantaneous power is uneven and the different subband signals are detected. Then the reverse decision is performed to detect positive positive results. The subband signal similar to the comb type signal is missed in the decision, and the parallel spectrum sensing.4. for multiple targets in the band is completed for the differences in the parameters of the participating nodes. In this paper, a more universal cooperative spectrum sensing algorithm is designed. The algorithm overcomes the local normalized spectrum uploading center and overcoming each node. The dynamic changes of noise level in time domain, space and frequency domain are accumulated, and then the test statistics are calculated with equal gain average or optimal weighted average in the fusion center, which can effectively eliminate the influence of high dynamic background noise on the spectrum sensing performance of the signal.5. design and hardware implementation of the spectrum sensor nodes. After completing the networking of the spectrum sensor, a high precision synchronous data acquisition method is designed. Finally, the single point and the cooperative perception measurement performance of the normalized spectral spectrum sensing algorithm and the energy spectrum sensing algorithm are compared.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN925

【相似文献】