稳定分布噪声中基于能量检测的频谱感知算法研究

发布时间：2018-05-03 00:20

本文选题：认知无线电 + 频谱感知　；参考：《大连海事大学》2017年硕士论文

【摘要】：随着无线通信技术的快速发展,有限的频谱资源变得越来越宝贵,传统的固定频谱分配方案已经满足不了日益增长的无线通信业务与频谱资源匮乏之间的矛盾。为了解决这一矛盾,认知无线电(Cognitive Radio,CR)作为一种新兴的技术应运而生,它使用一种"伺机"的方式接入和共享空闲频谱资源,有效地提高了频谱利用率。可靠的频谱感知是认知无线电实现其功能的关键技术之一,对其进行研究具有重要的应用价值。目前已经有许多成熟的信号检测技术可用于频谱感知,但在实际应用中,通信信道往往受到非高斯噪声的影响,使得传统基于高斯噪声模型假设下的检测算法不再适用。本文正是在这一背景下,采用α稳定分布作为脉冲噪声的数学模型,结合传统的能量检测(Energy Detection,ED)技术,提出了适用于脉冲噪声环境的三种改进的能量检测频谱感知算法。虽然α稳定分布是描述通信信道中非高斯噪声的最佳模型,但由于α稳定分布噪声不存在有限的二阶统计量,因此基于二阶统计量的能量检测算法在该噪声条件下性能严重退化甚至失效。为了解决这一问题,本文首先提出了基于非线性预处理的能量检测算法,通过将接收到的信号进行预处理变换使其具有二阶统计量,再进行能量检测,文中定义了保持抑制比参数来衡量非线性预处理函数的能力,仿真实验证明,该算法的检测性能要优于基于分数低阶统计量的检测算法和能量检测算法,并且非线性预处理函数的保持抑制比越大,其对应的检测性能越好;其次,基于核函数理论提出了一种核能量检测算法,其核心是通过核映射的方法将接收信号映射到变换域上,然后求变换域上的能量进行检测,仿真实验表明该算法在高斯噪声和α稳定分布噪声中都具有良好的检测性能;最后,针对调制信号具有循环平稳性的特点,提出了一种基于分数低阶循环谱能量的检测算法,仿真实验表明该算法在α稳定分布噪声中对调制信号的检测性能要优于基于二阶循环谱能量的检测算法和基于分数低阶统计量的检测算法。
[Abstract]:With the rapid development of wireless communication technology, the limited spectrum resources become more and more valuable. The traditional fixed spectrum allocation scheme can not meet the contradiction between the growing wireless communication services and the lack of spectrum resources. In order to solve this contradiction, Cognitive Radio (CR) as a new technology emerged as the times require. It uses a "opportunistic" way to access and share the idle spectrum resources, which effectively improves the spectrum efficiency. Reliable spectrum sensing is one of the key technologies to realize the function of cognitive radio. At present, there are many mature signal detection techniques that can be used in spectrum sensing, but in practical applications, the communication channel is often affected by non- noise, which makes the traditional detection algorithm based on Gao Si noise model is no longer applicable. Under this background, using 伪 stable distribution as the mathematical model of impulse noise and combining with the traditional energy detection technique, this paper presents three improved spectrum sensing algorithms suitable for impulse noise environment. Although the 伪 stable distribution is the best model for describing the non- noise in the communication channel, there is no finite second-order statistic for the 伪 stable distribution noise. Therefore, the energy detection algorithm based on second-order statistics has serious degradation and even failure under the noise condition. In order to solve this problem, an energy detection algorithm based on nonlinear preprocessing is proposed in this paper. In this paper, the ability of measuring nonlinear preprocessing function by preserving rejection ratio parameters is defined. The simulation results show that the detection performance of this algorithm is better than that of the detection algorithm based on fractional low order statistics and the energy detection algorithm. And the larger the retention and suppression ratio of nonlinear preprocessing function, the better the detection performance. Secondly, a kernel energy detection algorithm based on kernel function theory is proposed. The core of the algorithm is to map the received signal to the transform domain by the kernel mapping method, and then calculate the energy in the transform domain to detect. The simulation results show that the algorithm has good detection performance in both Gao Si noise and 伪 stable distributed noise. A detection algorithm based on fractional low order cyclic spectrum energy is proposed for modulation signal with cyclic stability. Simulation results show that the proposed algorithm performs better than the second order cyclic energy detection algorithm and the fractional low order statistic algorithm in detecting modulation signals in 伪 stable distributed noise.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN925

【相似文献】