当前位置: 首页>博士论文>资源详情
随机共振理论研究及其在微弱信号检测中的应用
中文摘要

随着非线性理论的发展,随机共振理论为强噪声背景下微弱信号检测开辟了新的研究方向。与传统检测方法不同,随机共振是将微弱信号在噪声的“助推”下得到增强,以大幅值跃迁的形式凸显出来,实现微弱信号的检测。本课题围绕随机共振理论及其在微弱信号检测中的应用展开深入细致地研究,摒弃了传统随机共振检测思路,以随机共振系统非线性动力学行为作为切入点,基于系统吸引子曲线分析系统参数对产生随机共振现象的影响,提出了基于吸引子曲线的调参随机共振理论,为微弱信号的检测提供了新方法,并且拓宽了利用随机共振理论对微弱信号检测的应用领域。 首先,针对经典随机共振系统检测微弱信号时受小频率参数条件限制的问题,引入“阻尼系数”这一参数,论证了阻尼系数对大参数微弱信号检测的影响,提出了基于吸引子曲线的调参随机共振检测大参数微弱信号的方法。该方法通过调整阻尼系数可实现大参数微弱信号的检测,联合调整系统参数能够更好地识别出微弱信号的频率特征,即使在低采样率的情况下仍具有较强的稳定性,并以Kramers跃迁速率为分析手段,将基于吸引子曲线的随机共振与经典随机共振相契合。同时将提出的方法应用于微弱无线电信号的检测,设计了随机共振高频无线电微弱信号检测系统。 其次,针对单随机共振系统检测结果信噪比低,对于无线电信号来说,无法满足解调或者解码要求的问题,提出了三级级联调参随机共振加强的方法。该方法通过分别为每一级系统分配主控参数,使得级联随机共振系统的输出性能得到明显地提高。同时基于该加强方法设计了自拟合级数的级联随机共振检测系统,根据对检测结果的要求,选定级联的级数,避免资源浪费,提高检测效率,并且有针对性地设置级联系统参数,可控制系统输出,提高输出信噪比,与其他增强输出性能的方法相比,该方法的增强效果明显。 最后,针对常用自适应随机共振在参数寻优后期盲目搜索的问题,将基于透镜原理的反向学习策略引入到人工鱼算法中,提出了基于知识的改进人工鱼群自适应随机共振检测方法。寻优中对个体进行“微观”和“宏观”调整,并将系统结构参数与系统发生随机共振的关系条件作为知识指导算法寻优,该方法参数寻优效率高,能以更快的速度获取最优系统参数,提高系统检测性能,同时为了解决检测淹没在强噪声中频带宽的多频微弱信号的难题,提出一种并联式自适应随机共振检测方法,为更优地检测出频带宽的多频微弱信号提供了方法。 关键词:吸引子曲线;随机共振;参数调节;微弱信号检测;人工鱼群算法

英文摘要

With the development of the nonlinear theory, stochastic resonance (SR) becomes a new research direction for weak signal detection on the background of strong noise. Compared with the traditional detection methods, the SR system can achieve the detection of the weak signal by enhancing the weak signal in the boosting of the noise and highlighting it in the form of large-scale transition. The dissertation focuses on the theory of SR and its application in weak signal detection. Instead of the traditional detection idea of SR system, the nonlinear dynamic behavior of SR system is used as the starting point. The influence of system parameters on the generation of SR is studied with the attractor curve of the system. The SR theory based on attractor curve is proposed, which provides a new method for weak signal detection and broadens the application field of weak signal detection by the principle of SR. Firstly, for dealing with the problem that classical stochastic resonance system is subject to the restriction of small frequency parameter in weak signal detection, the parameter of damping coefficient is introduced, and the influence of the damping coefficient on the detection of the weak signal of the large parameter is demonstrated, a SR detection method based on attractor curve is presented, which is suitable for high frequency weak signal detection. Adjusting the damping coefficient can realize the detection of weak signal of large parameters, and combined with adjusting the system shape parameters can identify the frequency characteristics of weak signal better, even at the low sampling rate, it still has strong stability. Moreover, the SR based on the attractor curve is fitted with the classical SR by Kramers rate. In addition, the method is applied to the detection of weak radio signals, a detection system of the SR is designed, which can detect the weak radio signal of the high frequency. Secondly, due to the signal-to-noise ratio (SNR) of the system output is low and can not meet the requirements of demodulation or decoding for a radio signal when the detection of weak signal in a single SR system, an enhanced method of three-level cascaded stochastic resonance is proposed. By adjusting the main control parameter of each level, the output performance of the cascaded stochastic resonance system is improved obviously. At the same time, a self-fitting cascaded stochastic resonance detection system is designed based on the enhanced method. According to the requirements of the test results, the cascade series are selected to save process and improve the detection efficiency. Using the proposed enhanced method, it can control the respond of the system and improve the SNR of the output by setting parameters purposely. Compared with other methods to enhance the output performance, the enhancement effect of this method has obvious advantages. Finally, due to the shortcomings of blind searching at the later stage of parameter optimization in the common adaptive stochastic resonance (ASR) system, an ASR detection method with knowledge-based improved artificial fish swarm algorithm (AFSA) is proposed, it introduces the lens principle into the opposite process of the OBL (lensOBL), and achieves the micro and macro adjustment of the optimal individuals in order to enhance the global search ability. In addition, the relationship between the system structure parameters and the SR phenomenon is as the knowledge to guide the optimization of the algorithm. The proposed detection method enhances the convergence speed and the accuracy of optimization, and improves the detection performance of the system. Furthermore, in order to solve the difficulty of detecting the multi-frequency weak signal which submerged in the strong noise and had wide frequency bandwidth, the parallel ASR detection systems is proposed, it provides a method for detecting the multi-frequency weak signal optimally with the wide frequency bandwidth. Key words: attractor curve; stochastic resonance; parameter-tuning; weak signal detection; artificial fish swarm algorithm

作者相关
主题相关
看过该书的人还在看哪些书