【作者】 吴利平；

【导师】 李建东； 李赞；

【作者基本信息】 西安电子科技大学 ， 通信与信息系统， 2011， 博士

【摘要】 随着无线通信技术的飞速发展,通信信号的体制和调制样式复杂多样,频谱日益拥挤和重叠,导致背景噪声与干扰显著提高,电磁环境极其复杂。这种复杂的电磁环境对于无论军事领域还是民用领域的无线通信系统都会形成严重的电磁噪声干扰,甚至会使通信联络中断,因而对于无线通信系统尤其是对于接收端的信号检测与估计提出了更高的要求和更严峻的挑战。由于非线性随机共振处理技术对于滤除干扰噪声和检测微弱信号具有较好效果,甚至可以利用噪声来增强有用信号,使得这一非线性现象在生物、化学、电子、图像处理等领域得到广泛应用。尽管如此,将随机共振理论应用于无线通信系统,仍然面临模型设计、复杂信号处理、参数优化与自适应控制等诸多关键问题。因此,针对当前无线通信面临的迫切需求和严峻挑战,本文将非线性科学领域的随机共振技术引入到复杂电磁环境下的信号检测与估计当中,在深入研究典型随机共振系统原理及其特性,并解决随机共振相关瓶颈问题的基础上,提出基于非线性随机共振系统的微弱信号检测、非高斯信号接收和信号参数估计等关键技术,进一步提升无线通信系统的接收性能。首先,对典型非线性系统中的随机共振现象及其原理进行详细描述和深入研究,为后续随机共振的实际应用提供了理论基础。将周期信号和噪声输入双稳态模型和单阈值模型这两种典型随机共振系统,研究和分析了因非线性系统、信号和噪声之间的协同效应而产生的随机共振现象及其机理。提出了基于线性变换的系统参数调整策略,并解决了SR系统参数最优化和复杂信号处理等问题,获得了最佳的随机共振处理性能。接下来,提出了复杂电磁环境下基于双稳态SR系统的能量检测算法,证明了随机共振处理能够实现噪声能量向信号能量转移,从而有效削弱了噪声不确定度的影响,进而改善1-2dB检测性能。并且,提出了基于双稳态SR系统的截尾序贯检验算法,推导了最佳截尾门限,进一步缩短了50%的检测时间。此外,提出了基于广义SR系统的能量检测算法,推导了最佳SR噪声的概率密度解析式,仅通过增加SR噪声就能改善检测性能。其次,将随机共振系统应用到广义高斯噪声情况下的信号最佳接收中,从而降低了误码率和改善接收性能。复杂环境下背景噪声的非高斯分布特性会导致线性最佳接收算法性能的急剧恶化,因此在分析了阵列阈值随机共振系统输出信号特征基础上,提出了基于阵列阈值系统的非线性接收算法。采用基于直接搜索法的阈值噪声自适应调整策略,实现了输出信噪比和误码率的最优化。理论推导和仿真结果表明在拉普拉斯型等非高斯分布噪声下,所提算法的接收性能优于基于匹配滤波的线性最佳接收算法1-3dB,从而保障了非高斯噪声下无线通信系统的接收性能。最后,针对复杂环境下的直接扩频信号估计需求,提出了基于广义随机共振系统和特征分解技术的信号参数估计算法。发现了电平通过率估计算法中的随机共振现象,并提出了基于该SR模型的最大多普勒频移估计算法,通过实现背景噪声与多普勒估计器之间的匹配,改善多普勒估计性能达1.5dB以上。与此同时,根据直接扩频信号相似性和特征分解原理,提出了一种基于平均互相关和特征分解的伪码序列估计算法,性能分析和仿真结果表明所提算法有效克服了部分反码问题和提高了2dB以上的抗噪声性能。更多还原

【Abstract】 With the rapid development of modern wireless communication, communication systems and modulation schemes is gradually complicated and diversified. Spectrum was increasingly congested and overlapping, which lead to an increase of background noise and interference significantly. The electromagnetic environment is extremely complex. This complex electromagnetic environment has badly restricted transmission quality and reception performance of communication system on civil and military fields, which even caused connection broken suddenly. Thus, it offers higher demands and severer challenges for wireless communication, especially signal detection and estimation in communication terminal. Due to the nonlinear stochastic resonance (SR) technology has good effects in noise suppression and weak signal detection, it is widely applied in many different areas such as biology, chemistry, electronics, image processing. However, the application of stochastic resonance in communications is still facing many key problem including model design, complicated signal process, parameter optimization, adaptive control, et al. Therefore, on the basis of in-depth research on typical SR systems and solution to their belated bottle-neck problems, this thesis proposed some important technologies about weak signal detection, and parameter estimation based on SR, which further promoted the reception performance of wireless communications systems under complex electromagnetic environment.First, the SR phenomenon and its mechanism in typical systems are researched and analyzed in details, which provides the theorotic basis for its subsequent actual applications. Then, the periodic signal and noise are put into bistable stochastic resonance (BSR) and suprathreshold stochastic resonance (SSR) mode, and the SR phenomena caused to synergistic effect between signal, noise and nonlinear systems and their mechanism are research and analyzed. Moreover, the adaptive parameters adjust strategy based on linear transformation is proposed, and system parameter optimization and complex signal process problem are solved for optimal performance in SR systems.Next, nonlinear SR process technology is applied in weak signal detection, to improve detection performance and reduce detection time. An energy detection (ED) based on BSR is proposed under the complex environment. By utilizing the SR process of received signal, proposed algorithm can achieve energy transfer from noise to signal, thus effectively eliminates the influence of interference noise and improves 1～2dB detection performance. Then, a truncated sequential probability ratio test (SPRT) algorithm based on BSR is introduced and the optimal truncated threshold is derived to further shorten detection time about 50%. Moreover, we propose an ED algorithm based on generalized stochastic resonance (GSR), which can improve the detection performance by adding SR noise with special probability density function.Then, nonlinear SR system is applied in signal reception under Generalized Gaussian noise conditions, to reduce bit error rate (BER) and improve reception performance. Because non-Gaussian interference noise brings to performance degradation of linear best receiving algorithms designed under Gaussian noise, on the basis of analysis on the output signal feature of SSR system, a nonlinear receiving algorithm is proposed based on SSR. Then, a strategy for parameters adaptive adjustment is put forward to ensure the optimization of output SNR and BER. Theory deduction and simulation results show that under non-Gaussian interference noise, e.g. Laplacian noise, the received performance of proposed algorithm is superior to linear best receiving algorithm based on matched filter (MF) for 1-3dB, thus it guarantees the reception performance of wireless communications systems under non-Gaussian noise.Finally, according to parameter estimation demand of spread spectrum (DS) signal, we propose the maximum Doppler shift estimation based on GSR and the PN estimation based on eigenanalysis. The SR phenomenon is discovered in Doppler shift estimation, and new Doppler shift estimation is proposed based on this SR mode. Through a low-pass filtering processing of the received signal, interference noise and the Doppler estimator can be matched, thus the estimation performance of the maximum Doppler shift can be improved more than 1.5dB. Meanwhile, a blind estimation for PN sequence is proposed based on the similarities among the DS signals and the eigenanalysis technique. Simulation results show that compared with the existing algorithms, the proposed algorithm not only overcomes the partial-encode problem, but also improves estimation performance by 2dB. Thus, it achieves an accurate estimation of PN sequence in low SNR conditions.更多还原