Advanced Detection Schemes of Digital Signals in Impulse Noise
Advanced Detection Schemes of Digital Signals in Impulse Noise
复制标题
脉冲噪声中数字信号的先进检测方案
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
Khodr A. Saaifan
中科院分区：
文献类型：
--
作者：
Khodr A. Saaifan
在本文中，我们研究了减轻无线通信信道中脉冲噪声影响的最佳方法。首先，我们提出了一个测量活动来验证米德尔顿a类（MCA）脉冲噪声模型的统计特性。这项运动测量的是破坏2.4 GHz工业、科学和医疗（ISM）频段的无线干扰。我们将这一验证扩展到表征多天线系统脉冲噪声的空间耦合和相关性。然后，我们研究了被MCA噪声破坏的二进制信号的最佳检测器。我们将MCA模型近似为单个加权高斯密度，使得最优检测器的非线性可以用封闭形式表达式来评估。近似MCA模型使用阈值检测方案区分高斯和脉冲事件的噪声概率密度函数（PDF）。通过这种近似，我们对最优非线性在减小脉冲噪声影响方面的行为提供了精确的分析。我们还引入决策边界分析来证明和分析在不同MCA噪声环境下最优检测器的性能。我们还近似非线性地研究了其他次优检测器的行为，如局部最优检测器（LOD）和裁剪检测器。作为次优方法，我们使用线性段进一步逼近最优非线性，引入新的次优检测器，如分段线性检测器和类裁剪检测器。其次，我们扩展了近似MCA模型，得到了存在衰落和脉冲噪声时的时空分集的最佳组合方案。我们假设完全了解噪声状态，以评估最佳组合方案在瑞利衰落和MCA噪声下的时间、接收和发射/接收分集的分析性能。虽然这个假设是不现实的，但它导致了脉冲噪声下最优组合方案的性能界很紧。这些评估使我们能够研究相对于发射和接收天线数量的空间分集性能损失。最后，我们在正交频分复用（OFDM）系统的脉冲噪声抑制问题中利用了频谱维数。本文首先从存在MCA噪声的衰落信道中OFDM系统的接收机设计入手，包括最优接收机和传统OFDM检测器的设计。然后，我们评估了在有MCA噪声的平坦衰落条件下的最佳接收机的性能上限。这一分析表明，在脉冲噪声方面，最优检测器比传统的OFDM检测器有显著的性能改进。最后，我们提出了一种结合稀疏贝叶斯学习（SBL）的球解码算法，以实现在MCA噪声下的最优OFDM检测器。
In this thesis, we investigate the optimum approach to mitigate the effects of impulse noise in wireless communication channels. First, we present a measurement campaign to verify the statistical properties of a Middleton Class-A (MCA) model for impulse noise. This campaign measures wireless interference that corrupted a 2.4 GHz industrial, scientific, and medical (ISM) band. We extend this verification to characterize the spatial coupling and correlation of impulse noise for multiple antenna systems. We then investigate the optimum detector for binary signals corrupted by MCA noise. We approximate the MCA model to a single weighted Gaussian density such that the nonlinearities of the optimum detector can be evaluated in a closed-form expression. The approximate MCA model discriminates the noise probability density function (PDF) of Gaussian and impulsive events using a threshold detection scheme. By means of such approximations, we provide a precise analysis for the behaviors of optimum nonlinearities in reducing the effects of impulse noise. We also introduce a decision boundary analysis to justify and analyze the performance of the optimum detector in different MCA noise environments. We also approximate the nonlinearities to investigate the behaviors of the other suboptimum detectors such as a locally optimum detector (LOD) and a clipping detector. As a suboptimum approach, we further approximate the optimum nonlinearities using linear segments to introduce new suboptimum detectors such as a piecewise linear detector and a clipping-like detector. Next, we extend the approximate MCA model to derive the optimum combining schemes for time and space diversity in the presence of fading and impulse noise. We assume perfect knowledge of noise states to evaluate the analytical performances of the optimum combining schemes for time, receive, and transmit/receive diversity in Rayleigh fading and MCA noise. Although this assumption is unrealistic, it leads to a tight performance bound for the optimum combining schemes in impulse noise. These evaluations allow us to study the performance loss of spatial diversity with respect to the number of transmit and receive antennas. Finally, we utilize the spectral dimensions in the mitigation problem of impulse noise for orthogonal frequency division multiplexing (OFDM) systems. We start with the receiver design of OFDM systems applied to fading channels with MCA noise, such as the optimum receiver and a conventional OFDM detector. We then evaluate an upper performance bound for the optimum receiver in flat fading with MCA noise. This analysis indicates a significant performance improvement of the optimum detector over a conventional OFDM detector in impulse noise. We finally develop a sphere decoding along with sparse Bayesian learning (SBL) to realize the optimum OFDM detector in MCA noise.