基于非线性随机共振的弱信号检测理论研究

发布时间：2018-05-01 07:08

本文选题：低信噪比 + 微弱信号检测　；参考：《西安电子科技大学》2015年博士论文

【摘要】：随着无线通信设备数量日益庞大、信号调制方式更加复杂多样、无线频谱日益拥挤和重叠,无线通信系统的背景噪声与干扰显著提高,导致接收信号常常被淹没在强背景噪声中从而呈现出低信噪比的特点。这些因素将对无线通信系统在军事以及民用领域的应用产生极其恶劣的影响,严重的情况下会导致通信线路的中断,无线通信系统接收端的微弱信号检测正面临着严峻的挑战。因此,如何提升低信噪比条件下的微弱信号检测性能成为当前无线通信系统中亟需解决的关键问题。目前,传统的微弱信号检测方法通常是在线性系统框架下采用线性信号处理的方法,核心思想都是通过抑制噪声来提高微弱信号检测的性能,这些方法面临以下问题：1、抑制噪声的同时,被测信号也受到抑制或损失,而且在强背景噪声条件下,经过线性处理后的信号仍然不能满足检测灵敏度的要求,很难对被测信号进行有效地提取；2、线性系统本身也是产生噪声的源,在整个信号处理过程中,涉及的系统越多,产生额外噪声的概率也越大；3、线性系统本身不具备提升信噪比的功能,随着背景噪声的增加,系统的输出信噪比会相应地下降。为了保障无线通信系统的通信质量与传输性能,本文将非线性随机共振技术引入到通信信号的检测理论中,针对低信噪比条件下的信号波形检测、能量检测及图像增强等方面的问题展开研究。随机共振描述了一种奇特的非线性物理现象：在特定的条件下,.非线性系统、微弱信号和背景噪声三者之间达到匹配状态从而产生协同作用,此时无序的噪声能量向有序的微弱信号能量进行转移,从而使得微弱信号得到增强。随机共振技术的提出颠覆了人们通常认为噪声是有害因素的认识,将噪声从有害的因素变为有利于信号传输的因素,在低信噪比下的微弱信号检测中显示出独特的优势。本文针对低信噪比条件下微弱信号检测中出现的新需求和面临的新挑战展开研究工作,全文研究内容主要分为以下四个方面：1、在高斯背景噪声条件下,给出了基于双稳态随机共振系统的微弱周期信号、非周期二进制调制信号处理机制及性能的定量研究。针对双稳态系统输入信号为周期信号的情况,综合分析了周期信号频率和噪声强度对随机共振效应的影响,推导了双稳态系统参数的解析表达式,通过调节系统参数确保了周期随机共振现象的产生。在此基础上,进一步定量分析了双稳态系统的信噪比增益。针对双稳态系统输入信号为非周期二进制调制信号的情况,分析了双稳态系统输出响应机制,研究了双稳态系统响应速度和码元周期及调制频率之间的定量关系,推导了双稳态系统参数的解析表达式,通过调节系统参数确保了非周期随机共振现象的产生。在此基础上,从系统输出信噪比和信号传输误码率等方面分析了基于双稳态随机共振系统的非周期二进制调制信号处理的性能。2、在广义高斯背景噪声条件下,提出了基于非线性阈值系统的非线性信号波形检测算法。该算法首先对接收到的信号经过非线性阈值系统进行处理,然后对非线性阈值系统的输出信号特征进行分析,最后根据最小平均错误概率准则计算得到基于非线性阈值系统的非线性信号波形检测算法的误码率表达式。仿真结果表明：在高斯背景噪声条件下,线性最佳检测算法的误码率性能优于本文所提算法；在拉普拉斯噪声(非高斯噪声,属于广义高斯噪声)条件下,本文所提算法的误码率性能优于高斯背景噪声假设下提出的线性最佳检测算法。3、为了提升低信噪比条件下非零均值信号采用能量检测(Energy Detection, ED)算法的检测性能,提出了基于广义随机共振系统的改进的能量检测(Improved Energy Detection, IED)算法。该算法首先对接收信号添加一个直流分量,并借助偏移系数确定添加直流分量的最优幅值,使其与信号中的直流产生广义随机共振；其次,对共振后的信号进行采样和能量累加得到检测统计量,然后根据最小平均错误概率准则确定最佳检测门限并与检测统计量进行比较从而做出判决；最后从错误概率和检测样本点数两个方面给出算法的性能分析。理论推导和仿真结果表明：在低信噪比条件下,采用IED算法的错误概率性能优于ED算法；在相同的错误概率条件下,IED算法所需的检测样本点数较ED算法显著减少。4、针对低峰值信噪比条件下二值图像增强的视觉效果及性能需求,提出了基于双稳态随机共振系统的二值图像增强算法。该算法通过对二值图像的像素点按行或者列的方向进行扫描,将二维图像的像素点转换为一维非周期二进制脉冲振幅调制(Binary Pulse Amplitude Modulated, BPAM)信号,然后采用双稳态随机共振系统对BPAM信号进行增强,最后再将增强后的信号转换为二维图像的像素点从而得到增强后的二值图像。仿真结果表明,本文所提算法的二值图像视觉增强效果明显优于传统的采用中值滤波、维纳滤波以及数学形态学的图像增强方法。当所处理图像的峰值信噪比为7.31dB时,采用中值滤波、维纳滤波以及数学形态学的方法对峰值信噪比分别可以提升4.96dB、2.96dB和2.54dB,而采用本文所提算法对峰值信噪比可以提升11.14dB,明显优于传统的二值图像增强算法。
[Abstract]:With the increasing number of wireless communication devices, the modulation mode of the signal is more complex and diverse, the wireless spectrum is increasingly crowded and overlapped. The background noise and interference of the wireless communication system are greatly improved, and the reception signals are often flooded in the strong background noise to show the characteristics of low signal to noise ratio. These factors will be used in the wireless communication system. The application of military and civil fields has a very bad influence. In serious cases, the communication lines will be interrupted. The weak signal detection of the receiver in the wireless communication system is facing a severe challenge. Therefore, how to improve the performance of the weak signal detection under the condition of low signal to noise ratio has become an urgent need to be solved in the wireless communication system. At present, the traditional method of weak signal detection is usually using linear signal processing method under the framework of linear system. The core idea is to improve the performance of weak signal detection by suppressing noise. These methods are faced with the following problems: 1, when noise suppression is suppressed, the measured signals are also suppressed or lost, and at the same time, the signal is also suppressed or lost. Under the condition of strong background noise, the signal after linear processing still can't meet the requirement of detection sensitivity, it is difficult to extract the measured signal effectively. 2, the linear system itself is also the source of noise. In the whole process of signal processing, the more the system involved, the greater the probability of generating extra noise; 3, linear system base In order to ensure the communication quality and transmission performance of the wireless communication system, the nonlinear stochastic resonance technology is introduced to the detection theory of the communication signal to detect the signal waveform under the condition of low signal to noise ratio. Random resonance describes a strange nonlinear physical phenomenon: under specific conditions, the nonlinear system, the weak signal and the background noise reach the matching state between the three, and the disordered noise energy goes to the ordered weak signal energy at this time. The proposed random resonance technique subverts the understanding that noise is a harmful factor, and changes noise from harmful factors into factors that are beneficial to signal transmission, and shows unique advantages in weak signal detection under low signal to noise ratio. The research work of new demand and new challenge in signal detection is carried out. The main content of the full text is divided into four aspects: 1, under the condition of Gauss background noise, the weak periodic signal based on bistable stochastic resonance system and the quantitative study of the processing mechanism and performance of the non periodic binary modulation signal are studied. The input signal of the steady state system is a periodic signal. The influence of the periodic signal frequency and the noise intensity on the random resonance effect is synthetically analyzed. The analytic expression of the bistable system parameters is derived. By adjusting the system parameters, the periodic random resonance phenomenon is ensured. On this basis, the bistable system is further analyzed. The signal to noise ratio gain. Aiming at the non periodic binary modulation signal of the bistable system input signal, the output response mechanism of the bistable system is analyzed. The quantitative relation between the response speed of the bistable system and the code period and the modulation frequency is studied. The analytic expression of the bistable system parameters is derived, and the parameters of the system are adjusted by adjusting the system parameters. On this basis, the performance.2 of the non periodic binary modulation signal processing based on the bistable stochastic resonance system is analyzed on the basis of the output signal to noise ratio and the signal transmission error rate of the system, and the nonlinear threshold system based on the generalized Gauss background noise is proposed. This algorithm first deals with the received signal through a nonlinear threshold system, then analyzes the output signal characteristics of the nonlinear threshold system. Finally, the error rate table of the nonlinear signal waveform detection algorithm based on the nonlinear threshold system is calculated according to the minimum mean error probability criterion. The simulation results show that the error rate performance of the linear optimal detection algorithm is superior to the proposed algorithm under the Gauss background noise condition. Under the condition of Laplasse noise (non Gauss noise, belonging to generalized Gauss noise), the performance of the proposed algorithm is better than the linear optimal detection proposed under the Gauss background noise hypothesis. Algorithm.3, in order to improve the detection performance of the Energy Detection (ED) algorithm for the non zero mean signal under low signal to noise ratio, an improved energy detection (Improved Energy Detection, IED) algorithm based on the generalized stochastic resonance (SR) system is proposed. The algorithm first adds a DC component to the received signal and uses the offset system. The number is determined to add the optimal amplitude of the DC component to generate the generalized random resonance of the direct current in the signal. Secondly, the sampling and energy accumulation of the signals after the resonance are sampled and the detection statistics are obtained. Then, the optimal detection threshold is determined according to the minimum mean error probability criterion and is compared with the test statistics to make a decision. The performance analysis of the algorithm is given from two aspects of error probability and detection sample point. The theoretical derivation and simulation results show that the error probability performance of the IED algorithm is better than the ED algorithm under the condition of low signal to noise ratio. Under the same error probability condition, the number of detection sample points required by IED algorithm is significantly lower than that of the ED algorithm, and the number of detection points is significantly lower than that of the ED algorithm. A two value image enhancement algorithm based on bistable stochastic resonance system is proposed for two value image enhancement in the condition of peak signal to noise ratio (peak signal to noise ratio). By scanning the pixels of the two value image in the direction of row or column, the pixel point of the two-dimensional image is converted to one dimension non periodic binary pulse amplitude modulation. The Binary Pulse Amplitude Modulated (BPAM) signal is made, and then the BPAM signal is enhanced by the bistable stochastic resonance system. Finally, the enhanced signal is converted to the pixel point of the two-dimensional image and the enhanced two value image is obtained. The simulation results show that the two value image enhancement effect of the proposed algorithm is obviously better than that of the proposed algorithm. Traditional median filtering, Wiener filtering and mathematical morphology are used for image enhancement. When the peak signal to noise ratio of the processed images is 7.31dB, median filtering, Wiener filtering and mathematical morphology can be used to improve the peak signal to noise ratio (4.96dB, 2.96dB and 2.54dB) respectively, and the peak signal to noise ratio (PEO) ratio proposed in this paper is adopted. 11.14dB can be improved significantly better than the traditional two valued image enhancement algorithm.

【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】：2015
【分类号】：TN911.23

【相似文献】