基于并联自适应随机共振的微弱信号检测方法  

Method of weak signal detection based on parallel self-adaptive stochastic resonance

作　　者：张勇亮[1] 李国林[1] 张晓瑜[2] ZHANG Yong-liang LI Guo-lin ZHANG Xiao-yu(Department 7,Naval Aeronautical and Astronautical University, Yantai 264001, China Command Automation Station, Xinjiang Military Region, Urumqi 830042, China)

机构地区：[1]海军航空工程学院7系,山东烟台264001 [2]新疆军区指挥自动化站,新疆乌鲁木齐830042

出　　处：《计算机工程与设计》2017年第5期1324-1330,共7页Computer Engineering and Design

基　　金：国防预研基金项目(9140A27020214JB14435)

摘　　要：针对传统大参数信号自适应随机共振方法存在的变换尺度变化范围选取缺乏固定标准、参数自适应效率低、检测到的目标信号不够明显等不足,提出一种基于并联自适应随机共振的微弱信号检测方法,实现强噪声背景下大参数微弱信号的快速、有效检测。推导出基于采样频率的变换尺度的最大变化范围,将该范围平均分段,以输出信噪比为适应度函数,在变换尺度各子搜索范围和共振系统参数搜索范围内,采用带极值扰动的简化粒子群算法对变换尺度及系统参数进行自适应优化选择;将优化得到的变换尺度和系统参数分别作为并联各子随机共振系统的输入信号变换尺度和系统参数;将各子系统的输出响应分别进行自相关处理后,合成为最终的系统输出响应。仿真及实际应用结果表明,该方法对强噪声背景中的微弱目标信号具有准确的检测能力,能够有效提高参数自适应效率,突出检测到的目标信号,增强强噪声背景下微弱信号的检测能力。Aiming at the disadvantages of traditional self-adaptive stochastic resonance （SR） system for large parameters signals, such as no fixed standard to select the variation range of changeable scale, low efficiency of parameters self-adaption, unobvious target signals being detected, and so on, a method of weak signal detection based on parallel self-adaptive SR （PSASR） was pro-posed, which realized a fast and effective detection of the weak signals under conditions of large parameters in high strong noise background. The max variation range of the changeable scale was developed, the search range was divided into several parts equally, and the changeable scale in each search subrange, as well as the system parameters of SR system in their respective search range, were optimized using an improved PSO named extremum disturbed and simple particle swarm optimization （tsP- SO） , with SNR of output response as the fitness function. The changeable scales and parameters obtained through the optimiza-tion were respectively taken as the changeable scale of the input signal and the system parameters for each SR system connected in parallel. All the output responses of each SR sub-system were precessed by auto correlation analysis, and then composed as the final output responses of the system. Results of simulation and engineering application indicate that, the method proposed can accurately detect the weak target signals in high strong noise background, the efficiency of parameters self-adaption is im-proved effectively, the target signals are emphasized, and the detection ability of weak signals in high strong noise background is enhanced.

关 键 词：随机共振 自适应 并联 带极值扰动的简化粒子群算法 自相关分析 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]