基于三阶循环累积量的二维谐波信号的参数估计

本文从循环平稳的观点出发来研究乘性和加性噪声中的二维谐波参数估计问题.利用循环累积量的性质得到了二维三阶循环累积量单一记录估计的统计性质.在此基础上,提出了基于三阶循环累积量特殊切片的二维谐波分量个数和频率对的估计方法,并得到了估计的强相容性质和强收敛速度.仿真试验对算法作了说明.     

基于三阶累积量对角切片的直接自适应谱线增强算法

依据随机平方相位耦合信号三阶累积量不为零的特点，提出了从高斯噪声环境中恢复平方相位耦合信号的基于三阶累积量对角切片的直接自适应谱线增强算法。该算法恢复平方相位耦合信号、抑制高斯噪声的性能优于传统的自适应谱线增强算法。仿真结果证实了这种算法的有效性。     

基于残差三阶累积量的机动检测新方法

针对在有高斯量测噪声情况下目标机动检测问题,提出了一种基于残差三阶累积量的 机动检测算法,通过利用Kalman滤波实时产生残差数据而进行的三阶累积量的计算,抑制了 高斯噪声对目标机动的影响,且实现了机动检测的实时性。仿真实验结果验证了该算法的有 效性和稳健性。     

一种新的盲解卷积准则和算法

定义了盲解卷积问题的期望解后 ,将二阶累积量和四阶累积量合并为一个新的统计量 ,称为归一化累积量 ,考察信号通过线性时不变系统时归一化累积量的特性 ,形成一个基于归一化累积量的盲解卷积准则 ;借助于经典的最陡梯度算法 ,导出了一种新的盲均衡算法 ,计算机模拟验证了该算法     

基于输出累积量的ARMA（AR0系统的一致性阶次递推辩识

本文提出了一种根据系统输出的观测数据对ＡＲＭＡ（ＡＲ）系统进行盲识别的新算法。该模型由独立同分布非高斯随机序列驱动，其输出序列中含方差未知的加性高斯噪声。通过求解基于三阶累积量谱的代价函数。该算法以模型阶次递推形式同时辩识ＡＲＭＡ的系统阶次和估计出系统能数。     

基于高阶循环累积量的QAM载波频率估计

利用通信信号的循环平稳特性,通过对QAM信号高阶循环累积量的研究,提出了一种基于高阶循环累积量的星形和方形QAM载波估计算法。并且对该算法进行了理论分析,仿真结果证明,该算法的正确性和有效性。     

基于WCPO准则的鲁棒四阶累积量盲波束形成算法

基于边界误差性能最优准则(WCPO)的鲁棒性方法是波束形成算法领域内的最新研究热点,本文将该鲁棒性方法与基于高阶累积量的盲波束形成算法相结合,提出了一种新的改进鲁棒性的四阶累积量盲波束形成算法,即WCPO-RCUM算法.该新算法相较于采用传统对角线加载技术的高阶累积量盲波束形成算法RCUM而言,进一步提高了稳健性和输出信干噪比.计算机仿真验证了该算法的可行性和有效性.     

基于高阶累积量的目标机动检测新方法

机动检测是多模型目标跟踪中的一个关键问题.在卡尔曼滤波中,当目标机动被噪声淹没时,传统的机动检测算法将失效,多分辨方法虽然能够有效地抑制噪声,可靠检测机动,但由于计算复杂导致严重的检测延迟,从而限制了它的应用.本文提出一种基于三阶累积量的机动检测新算法,它有效地克服了上述二者的缺陷.由于高阶累积量能够抑制高斯噪声,因此在三阶累积量域易于检测机动.同时通过采用逐点更新法,可实时进行机动检测.仿真结果表明,该算法优于传统算法和多分辨方法,特别是在低信噪比的情况下.     

高阶累积量盲均衡算法的优化设计

基于高阶累积量盲均衡算法的优化设计,提出了一种新的盲均衡算法。该算法通过引入优化思想,利用模拟退火算法来对高阶累积量的SW准则进行优化;利用模拟退火算法的全局收敛性以及快速收敛性,提高了算法的性能。计算机仿真结果表明,该算法具有较好的收敛性能及抗误码性能。     

基于输出累积量的 ARMA(AR)系统的一致性阶次递推辩识(英文)

本文提出了一种根据系统输出的观测数据对ARMA（AR）系统进行盲识别的新算法。该模型由独立同分布非高斯随机序列驱动，其输出序列中含方差未知的加性高斯噪声。通过求解基于三阶累积量谱的代价函数，该算法以模型阶次递推形式同时辩识ARMA的系统阶次和估计出系统参数。文章给出了该算法一致收敛性的证明，并对两类不同阶次的最小及非最小相位ARMA系统的AR参数及阶次辩识进行了数字仿真，结果令人满意。     

Nouvelles approches ď estimation des paramètres autorégressifs variables utilisant les cumulants

In this paper, we present two new cumulant-based methods for time-varying ar parameter estimation : a batch-type evolutive method and an adaptive gradienttype algorithm. The evaluation of these techniques is performed through simulations on synthetic signals in free-noise case and when data are corrupted by an additive, zero-mean, Gaussian white noise. We compare them to their autocorrelation-based counterparts. The obtained results show, when using an appropriate criterion, the superiority of the cumulant-based evolutive method over both its autocorrelation-based counterpart and the cumulant-based gradient-type algorithm at expense of a great computational complexity.     

On blind beamforming for multiple non-Gaussian signals and theconstant-modulus algorithm

A new method for blindly separating multiple cochannel non-Gaussian signals received by a sensor array is presented. The method is based on a cumulant-based least-squares criterion that, for identically distributed negative-kurtosis signals, is proven to be identical to the “2-2” constant-modulus (CM) cost function commonly used by CM algorithms. A computationally simple algorithm is proposed to minimize the criterion. The algorithm performs well even when the number of samples is small, thus allowing its application in dynamic environments (e.g., moving emitters). For the special case of two signals only, the minimization is obtained analytically. Simulation results are included     

Extended-aperture angle-range estimation of multipleFresnel-region sources with a linear tripole array using cumulants

This paper presents a cumulant-based algorithm to achieve aperture extension for estimating the directions-of-arrival (DOAs) and the ranges of multiple Fresnel-region sources using a linear tripole array. The proposed algorithm defines two cumulant-based matrices, from which the DOA and the range of each source are estimated from the source's tripole steering vector using the ESPRIT technique. These are then used as coarse reference estimates to disambiguate the cyclic phase ambiguities induced from the spatial phase factors when the inter-sensor spacing exceeds a half wavelength. The algorithm does not require two-dimensional searching or parameter pairing, and can resolve 3(L−1) sources with L tripoles. The extension of the proposed algorithm by formulating multiple cumulant matrices and using parallel factor (PARAFAC) analysis is also presented. Simulation results are provided demonstrating the significant improvement in the performance over that of several existing algorithms.     

Noise reduction in brain evoked potentials based on third-order correlations

In this paper, we use third-order correlations (TOC) in developing a filtering technique for the recovery of brain evoked potentials (EPs). The main idea behind the presented technique is to pass the noisy signal through a finite impulse response filter whose impulse response is matched with the shape of the noise-free signal. It is shown that it is possible to estimate the filter impulse response on basis of a selected third-order correlation slice (TOCS) of the input noisy signal. This is justified by two facts. The first one is that the noise-free EPs can be modeled as a sum of damped sinusoidal signals and the selected TOCS preserve the signal structure. The second fact is that the TOCS is insensitive to both Gaussian noise and other symmetrically distributed non-Gaussian noise, (white or colored). Furthermore, the approach can be applied to either nonaveraged or averaged EP observation data. In the nonaveraged data case, the approach therefore preserves information about amplitude and latency changes. Both fixed and adaptive versions of the proposed filtering technique are described. Extensive simulation results are provided to show the validity and effectiveness of the proposed cumulant-based filtering technique in comparison with the conventional correlation-based counterpart.     

Image motion estimation algorithms using cumulants

A class of algorithms is presented that estimates the displacement vector from two successive image frames consisting of signal plus noise. In the model, the signals are assumed to be either non-Gaussian or (quasistationary) deterministic; and, via a consistency result for cumulant estimators, the authors unify the stochastic and deterministic signal viewpoints. The noise sources are assumed to be Gaussian (perhaps spatially and temporally correlated) and of unknown covariance. Viewing image motion estimation as a 2D time delay estimation problem, the displacement vector of a moving object is estimated by solving linear equations involving third-order auto-cumulants and cross-cumulants. Additionally, a block-matching algorithm is developed that follows from a cumulant-error optimality criterion. Finally, the displacement vector for each pel is estimated using a recursive algorithm that minimizes a mean 2D fourth-order cumulant criterion. Simulation results are presented and discussed.     

基于累积量的两层前馈神经网络盲辨识

由于非线性系统输出是其参数的非线性函数,直接利用高阶累积量辨识两层前馈神经网络(FNN)通常是十分困难的。为解决这一问题,该文提出两种基于四阶累积量的FNN辨识方法。第一种方法,FNN的隐元在其输入空间利用多个线性系统近似,进而FNN利用一统计模型混合专家(ME)网络重新描述。基于ME模型,FNN参数可利用统计期望值最大化(EM)算法获得估计。第二种方法,为简化FNN的ME模型,引入隐含观测量。基于隐含观测量估计,FNN被分解为多个单隐元的训练问题,进而整体FNN可利用一两阶层ME描述。基于单隐元的参数估计,FNN可利用一具有更快收敛速度的简化算法获得估计。     

Parametric cumulant based phase estimation of 1-D and 2-Dnonminimum phase systems by allpass filtering

This paper proposes a parametric cumulant-based phase-estimation method for one-dimensional (1-D) and two-dimensional (2-D) linear time-invariant (LTI) systems with only non-Gaussian measurements corrupted by additive Gaussian noise. The given measurements are processed by an optimum allpass filter such that a single Mth-order (M⩾3) cumulant of the allpass filter output is maximum in absolute value. It can be shown that the phase of the unknown system of interest is equal to the negative of the phase of the optimum allpass filter except for a linear phase term (a time delay). For the phase estimation of 1-D LTI systems, an iterative 1-D algorithm is proposed to find the optimum allpass filter modeled either by an autoregressive moving average (ARMA) model or by a Fourier series-based model. For the phase estimation of 2-D LTI systems, an iterative 2-D algorithm is proposed that only uses the Fourier series-based allpass model. A performance analysis is then presented for the proposed cumulant-based 1-D and 2-D phase estimation algorithms followed by some simulation results and experimental results with real speech data to justify their efficacy and the analytic results on their performance. Finally, the paper concludes with a discussion and some conclusions     

AR modelling of skewed signals using third-order cumulants

Algorithms for selecting the order and estimating the parameters of an AR process, which is driven by noise having an underlying non-Gaussian distribution, from the observed noisy time series are presented. The order selection algorithm makes use of the growing memory covariance predictive least-squares (GMCPLS) criterion together with diagonal slices of the third-order cumulant plane. A triangular region of the third-order cumulant plane is used to estimate the model parameters. Extensive simulation results are presented and based on these trends, one of which has been verified using real data obtained from a rotating machine, recommendations are made on the efficacy of methods for AR order selection and parameter estimation problems     

A novel blind estimation algorithm

In this paper, we propose a cumulant-based blind signal estimation algorithm for estimating the channel matrix in an n-sensor m-source system. The only available information is the output of the n sensors. The algorithm first deduces the number of sources, which may be greater than or equal to the number of sensors, from the output cumulant matrix. Then, by suitably arranging the elements within that matrix, the entries of the original channel matrix are estimated row by row. Simulations results are given to illustrate the performance of the algorithm     

自适应滤波算法综述

分析了最小均方误差滤波和递归最小二乘滤波算法、自适应滤波的神经网络方法、基于QR分解的方法、统一模型下的自适应滤波及基于高阶累积量的自适应算法的优缺点,并对自适应滤波算法的未来发展做了展望。