一种非线性动力学降噪法应用于海底探雷抗混响的探讨

文章所述降噪法建立在背景信号能够用低维动力学系统建模的基础之上，利用待分离信号与干扰背景的频谱差异，通过连续拟合的两层RBF网来构造非线性逆滤波器以达到信号与背景分离的目的。本文给出了分别以仿真混沌信号和实测混响信号为背景与单频脉冲信号相分离的示例，证明了该算法在混响背景下实现信号分离的可行性(分离前后信混比提高了约16dB)，分析了算法有效使用所必需的频谱差异条件(不小于0．15个圆周频率)，为抑制混响提供了一条新的思路。     

基于相空间重构与独立分量分析的局部独立投影降噪算法

针对故障分析信号中存在噪声问题,提出一种将相空间重构与独立分量分析相结合的局部独立投影降噪算法。其中相空间重构的目的在于从高维相空间中恢复混沌吸引子,独立分量分析能够找到信号的主流形,选择邻域是为了将特征相近的相点结合在一起。使用该方法对正弦仿真信号和Lorenz仿真信号进行降噪处理,结果表明局部独立投影降噪算法的降噪效果与局部独立分量分析算法降噪效果接近,但优于全局投影降噪算法。运用该方法对低速重载轴承振动信号进行分析,准确判断出轴承故障。     

基于混沌理论的海底混响中目标回波提取

提出了基于混沌理论的混响中目标回波提取新方法。该方法主要得益于一种新的预测模型,该模型基于径向基函数神经网络,综合利用了时间序列的前向和后向预测,解释了该模型用于混沌信号分离的基本原理,用几种混沌时间序列分析了该模型用于混沌信号建模和谐波信号提取的性能。该方法用于湖试混响中目标回波提取的结果表明:该模型可以用于提取信混比不小于1dB的目标回波。     

主动声纳中目标回波的一种非线性提取方法

介绍了一种混沌时间序列的非线性动力学降噪方法——变尺度概率净化法,并将之优化应用于混响背景下目标回波的提取。该方法以近似条件下不含目标回波的混响数据作为参考,使目标回波分离后的混响背景数据与参考混响数据在相空间上具有一致的概率分布,从而实现目标回波的提取。在信混比不低于6dB的情况下,该方法具有较好的目标回波提取效果。文中对算法作了一些改进,降低了算法的运算量和运行所需计算机内存,而降噪性能并没有下降。     

基于混沌优化的板形信号模式识别的研究

针对传统的基于最小二乘法板形信号模式识别方法抗千扰能力差、精度低，神经网络识别方法在实际应用中效果不佳的问题，通过对板形信号和板形识别数学模型的分析，首先将板形信号模式识别过程转化为函数的优化问题。为提高板形信号识别的精度和速度，以勒让德正交多项式作为板形缺陷的基模式，用模糊识别理论与混沌优化方法对该函数进行优化求解。采用模糊理论作为初步识别，用以降低混沌优化的求解维数和缩小搜索空间，借助梯度下降法的思想对混沌优化的局部搜索能力进行改善，从而进一步提高了混沌优化对板形信号模式识别的识别速度和精度。     

基于LP与小波包的复杂机电系统监测序列混沌降噪方法

针对流程工业生产系统监测数据存在强噪声和混沌性的特点,提出了一种局部投影方法(Local Projection Method)与小波包方法相结合的信号降噪方法。该方法先利用局部投影方法从动力学系统嵌入流形的角度进行多次迭代降噪,并根据关联维数来判定迭代终止;再利用小波包方法从频率的角度进行降噪,抑制高频噪声的干扰,取得了较好的降噪效果。用Lorenz时间序列进行仿真验证,对仿真时间序列加入不同程度的噪声,对比分析小波包、局部投影与该方法降噪后的相空间、SNR值和最大Lyapunov指数,证明了该方法对于中高强度噪声具有更好的降噪效果。并将该方法用于某压缩机组的实际监测数据降噪,评估三种方法的降噪效果,进一步验证了该方法的优越性。     

混响背景下主动声纳接收信号的可分离性探讨

混响背景下主动声纳的接收信号是否具有可分离性是盲分离在主动声纳抗混响中获得应用的前提条件。在建立混响背景下的接收信号模型的基础上,通过模型中源信号间相关性的分析,提出了一种接收信号是否具有可分离性的判断准则：在绝大多数时延条件下相关系数很小的发射信号,其混响背景下的接收信号具有可分离性。通过主动声纳的两种发射信号（CW和LFM）的仿真和海试实验研究对这一准则的可行性进行了检验。对于符合可分离性判断准则LFM信号,实验结果验证了其混响背景下接收信号的可分离性。     

信号的小波尺度-频率表示及其在机械故障诊断中的应用

基于对信号的连续小波变换特点的分析，提出信号特征的小波尺度一频率表示方法。该方法对污染的脉冲信号和正弦信号的应用结果表明该方法能有效提取强噪声背景下的周期成分。以变速器齿轮的故障诊断为例，用该方法对变速箱齿轮振动信号的尺度-频率表示进行分析，结果表明该方法能非常有效地将不同磨损状况的齿轮振动信号的特征表现出来。     

基于EEMD和混沌的信号特征提取方法及应用

提出了一种基于集合经验模态分解(EEMD)和混沌相结合的信号特征提取方法,应用于婴儿呼吸信号哮喘检测中。EEMD把呼吸的局部信号分解成一系列频率从高到低的模态分量,对各分量与局部呼吸信号进行相关分析,并给出各分量的Hilbert谱,通过局部分析的结果初步判断婴儿是否患有哮喘;然后,以EEMD局部信号检测出来的信号频率作为混沌振子检测的频率,对全局呼吸信号进行整体检测及分析,由混沌的间歇周期可以得出原始呼吸信号的频率,准确确定婴儿哮喘诊断结果。对EEMD和混沌算法的应用存在的问题进行了改进,将其应用到实测信号的分析中,验证了方法的有效性。该方法能够正确地反映信息特征,准确率高。     

捷变正弦信号源波形建立时间的精确测量

针对正弦信号源捷变状态切换后的建立时间、开机后波形建立时间以及过载恢复时间等的精确测量问题,提出了一种基于局域波形四参数拟合的测量分析方法,然后将拟合模型参数拓展到全局,进而获得拟合回归波形与过渡过程波形的回归残差波形。该波形的收敛过程反映了正弦波建立过程中的残差收敛变化过程。以它为目标对象,加上主观设定的建立时间的条件判据,可以获得正弦建立时间的起始和终止两个时刻点,最终获得完整的正弦信号建立时间。在两组不同条件下的状态切换实验结果,验证了该方法的有效性和可行性。该方法也可以推广应用到脉冲调频、脉冲调幅、脉冲调相、捷变频信号的建立时间测量评价中。     

预白化与小波变换联合的混响抑制方法研究

针对主动声呐在强混响背景下匹配滤波器检测性能不理想的问题,提出了将预白化与小波变换联合处理方法。首先将混响预白化处理,然后通过小波变换将混响和回波信号分解到不同的尺度空间下,在尺度空间进行阈值量化处理,对有用的信号成分进行数据重构,最后利用重构的数据进行匹配滤波处理,使得混响被有效抑制,改善检测性能。仿真和试验数据处理结果验证了该方法的有效性,信混比可以提高约3 dB。     

Binaural Speech Separation Algorithm Based on Long and Short Time Memory Networks

Speaker separation in complex acoustic environment is one of challenging tasks in speech separation. In practice, speakers are very often unmoving or moving slowly in normal communication. In this case, the spatial features among the consecutive speech frames become highly correlated such that it is helpful for speaker separation by providing additional spatial information. To fully exploit this information, we design a separation system on Recurrent Neural Network (RNN) with long short-term memory (LSTM) which effectively learns the temporal dynamics of spatial features. In detail, a LSTM-based speaker separation algorithm is proposed to extract the spatial features in each time-frequency (TF) unit and form the corresponding feature vector. Then, we treat speaker separation as a supervised learning problem, where a modified ideal ratio mask (IRM) is defined as the training function during LSTM learning. Simulations show that the proposed system achieves attractive separation performance in noisy and reverberant environments. Specifically, during the untrained acoustic test with limited priors, e.g., unmatched signal to noise ratio (SNR) and reverberation, the proposed LSTM based algorithm can still outperforms the existing DNN based method in the measures of PESQ and STOI. It indicates our method is more robust in untrained conditions.     

冲击混响法溶洞探测初步试验

提供一种区别于通常地矿应用的测声时差的新的溶洞声学测试方法一冲击混响法。通过有无溶洞两个模型的声学试验,对测量数据进行时域分析和时频分析,得到混响衰减系数和混响持续时间两个参数,结果表明：两个模型的混响衰减系数和混响持续时间都存在明显差异,空心模型的混响衰减系数较实心模型小．而混响持续时间则比实心模型大。因而为方便快捷地判断有无溶洞的冲击混响法提供了试验依据。     

A Novel Method of Distance Measurement Based on Pulse Position Modulation and Synchronization of Chaotic Signals Using Ultrasonic Radar Systems

This paper deals with a novel method of transmission and receipt of a signal based on both the property of two chaotic systems generating the same chaotic signal when they are synchronized and the property of pulse position modulation (PPM) to be insensitive to the distortions of the transmission channel. The method is discussed in the context of ultrasonic radar systems, in which the transmitter and receiver, which consist of ultrasonic sensors, are near each other, and the received signal consists of the transmitted signal reflected by an obstacle. A reference sinusoidal signal is superimposed to a chaotic signal generated by a master chaotic system, and the whole signal is modulated according to the PPM method and transmitted by the sensor. The received signal is demodulated, and the demodulated signal forces a slave chaotic system to generate the chaotic signal embedded in it, which allows recovery of the sinusoidal signal by subtracting this chaotic signal from the demodulated echo. The difference of the phases of the reference sinusoidal signal and the recovered sinusoidal signal allows computation of the time of flight of the signal and, consequently, the distance of the radar system from the obstacle. The novel method is illustrated and tested by both simulation and experiments. The interference problem between the considered radar system and other radar systems ( crosstalk) is also addressed, and a solution is proposed to avoid it.     

Hidden data transmission using time delay for separating useful signals from masking oscillations

A new method of hidden data transmission based on the use of time delay for the separation of useful signals from masking noise-like (chaotic or stochastic) oscillations is described. Mathematical models involving a source of chaotic oscillations have been studied using numerical methods. The transmission of a masked non-encoded signal and the pulsed data transmission using 0/1 bit code are considered.     

Hyperbolic chaos generator based on coupled drift klystrons

A scheme of the generator of chaotic oscillations in the microwave band is proposed that is based on two drift klystrons coupled in a ring circuit. The first klystron doubles the frequency of the input signal and the second klystron mixes the second harmonic signal and the reference signal representing a periodic sequence of radio pulses filled with the third harmonic, followed by separation of the differential first harmonic signal. As a result, the transformation of the signal phase during the period of pulse repetition is described by a stretching map of the circle (Bernoulli map) and demonstrates chaotic behavior. Results of numerical calculations are presented that confirm the implementation of the proposed mechanism to ensure the generation of robust, structurally stable chaos in the system under consideration.     

单驱动源混沌测量方法

混沌运动的基本特征表现为对初值的敏感依赖性。利用混沌轨道随初值的变化 ,定义了轨道距离 ,通过测量混沌轨道的距离测得该初始值。但混沌系统也同样对干扰极度敏感 ,初值相同的轨道由于电路中的噪声会指数分离 ,从而又完全否定了混沌系统可应用于小信号的测量 ,问题的关键是怎样降低电路的噪声。提出了一种新的混沌测量方法 ,与已有的混沌测量电路相比 ,该方法采用一个压控恒流源向混沌振子提供充放电电流。实验结果表明 ,它能改善测量系统的线性度和稳定性