Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572485
A. Swami
We consider the problem of estimating the parameters of a linear system, when the observed output and the control input are corrupted by multiplicative noise. We show that the classical cross-correlation techniques may be used if the multiplicative noises have non-zero mean; in the zero-mean case, higher-order cross-moments and cumulants must be used. Parametric, non-parametric and adaptive estimators are developed.
{"title":"Input-output System Identification In The Presence Of Multiplicative Noise","authors":"A. Swami","doi":"10.1109/SSAP.1994.572485","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572485","url":null,"abstract":"We consider the problem of estimating the parameters of a linear system, when the observed output and the control input are corrupted by multiplicative noise. We show that the classical cross-correlation techniques may be used if the multiplicative noises have non-zero mean; in the zero-mean case, higher-order cross-moments and cumulants must be used. Parametric, non-parametric and adaptive estimators are developed.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116797974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572522
M. Fargues, R. Hippenstiel
We recently investigated various spectral-based classification schemes designed to separate wideband transient signals and compared their performances to those obtained using a back-propagation neural network implementation [2]. The spectral-based measures considered include the Bhattacharyya distance, the divergence, the normalized cross-correlation coefficient, and the modified normalized cross-correlation coefficient. Results showed that accurate classification may be obtained using spectral-based measures and that the performances compare, or are sometimes better, to those obtained using neural networks when the training data used to train the neural network is small. In this paper we investigate the robustness of the spectral measures and the neural network approximation classification schemes to white additive noise degradation in the testing sets. Results show that the spectral-based techniques are more robust when the testing sets are degraded with noise.
{"title":"Robust Spectral-Based Techniques for Classification of Wldeband Transient Signals","authors":"M. Fargues, R. Hippenstiel","doi":"10.1109/SSAP.1994.572522","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572522","url":null,"abstract":"We recently investigated various spectral-based classification schemes designed to separate wideband transient signals and compared their performances to those obtained using a back-propagation neural network implementation [2]. The spectral-based measures considered include the Bhattacharyya distance, the divergence, the normalized cross-correlation coefficient, and the modified normalized cross-correlation coefficient. Results showed that accurate classification may be obtained using spectral-based measures and that the performances compare, or are sometimes better, to those obtained using neural networks when the training data used to train the neural network is small. In this paper we investigate the robustness of the spectral measures and the neural network approximation classification schemes to white additive noise degradation in the testing sets. Results show that the spectral-based techniques are more robust when the testing sets are degraded with noise.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"110 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116744134","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572534
F. Claveau, S. Lord, D. Gingras, P. Fortier
A laser-based contactless displacement measurement system developed at the National Optics Institute is used for data acquisition to analyze the mechanical vibrations exhibited by vibrating structures and machines. The analysis of these vibrations requires a number of signal processing operations which include the determination of the system conditions through a classification of various observed vibration signatures and the detection of changes in the vibration signature in order to identify possible trends. This information is also combined with the physical characteristics and contextual data (operating mode, etc.) of the system under surveillance to allow the evaluation of certain characteristics like fatigue, abnormal stress, life span, etc., resulting in a high level classification of mechanical behaviours and structural faults according to the type of application. The aim of this paper is to introduce the problem, the instrumentation, and the requirements in terms of statistical signal processing.
{"title":"Mechanical Vibration Analysis Using an Optical Sensor","authors":"F. Claveau, S. Lord, D. Gingras, P. Fortier","doi":"10.1109/SSAP.1994.572534","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572534","url":null,"abstract":"A laser-based contactless displacement measurement system developed at the National Optics Institute is used for data acquisition to analyze the mechanical vibrations exhibited by vibrating structures and machines. The analysis of these vibrations requires a number of signal processing operations which include the determination of the system conditions through a classification of various observed vibration signatures and the detection of changes in the vibration signature in order to identify possible trends. This information is also combined with the physical characteristics and contextual data (operating mode, etc.) of the system under surveillance to allow the evaluation of certain characteristics like fatigue, abnormal stress, life span, etc., resulting in a high level classification of mechanical behaviours and structural faults according to the type of application. The aim of this paper is to introduce the problem, the instrumentation, and the requirements in terms of statistical signal processing.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"EM-33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126528228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572486
Huang-Lin Yang, Chong-Yung Chi
This paper presents a new cumulant based phase estimation method for linear time-invariant (LTI) systems with only non-Gaussian measurements contaminated by Gaussian noise. An optimum allpass filter is designed to process the given measurements such that its output has a maximum Mth-order (2 3) cumulant in absolute value. It can be shown that the system phase is equivalent to the negative value of the optimum allpass filter phase except for a linear phase factor. Some simulation results are provided to support the proposed phase estimation method.
{"title":"A New Cumulant Based Phase Estimation Nonminimum-phase Systems By Allpass","authors":"Huang-Lin Yang, Chong-Yung Chi","doi":"10.1109/SSAP.1994.572486","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572486","url":null,"abstract":"This paper presents a new cumulant based phase estimation method for linear time-invariant (LTI) systems with only non-Gaussian measurements contaminated by Gaussian noise. An optimum allpass filter is designed to process the given measurements such that its output has a maximum Mth-order (2 3) cumulant in absolute value. It can be shown that the system phase is equivalent to the negative value of the optimum allpass filter phase except for a linear phase factor. Some simulation results are provided to support the proposed phase estimation method.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114215427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572543
J. Ward, A. Steinhardt
Advanced airborne radars must perform target detection in the presence of interference and heavy clutter, Space-time adaptive processing (STAP) refers to a class of adaptive filtering techniques that simultaneously rocess the spatial signals from an antenna .array and d e temporal signals from multiple pulses an order to suppress both jammin and clutter. A reduceddimension suboptimum STh' architecture utilizing multi le dop ler filter banks on each element is suggestel Digrent methods for choosing the doppler filters are considered and a condition which yields minimum clutter r a d is derived. PRI-staggered postdop ler meets the condition and rovides both ezcellent per6rmance with few degrees orfreedom and the abili t y to maintain low adapted doppler sadelobes. Adjacent bin post-doppler re uires more de rees of freedom when low doppler sidelo%es are desire!
{"title":"Multiwindow Post-Doppler Space-Time Adaptive Processing","authors":"J. Ward, A. Steinhardt","doi":"10.1109/SSAP.1994.572543","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572543","url":null,"abstract":"Advanced airborne radars must perform target detection in the presence of interference and heavy clutter, Space-time adaptive processing (STAP) refers to a class of adaptive filtering techniques that simultaneously rocess the spatial signals from an antenna .array and d e temporal signals from multiple pulses an order to suppress both jammin and clutter. A reduceddimension suboptimum STh' architecture utilizing multi le dop ler filter banks on each element is suggestel Digrent methods for choosing the doppler filters are considered and a condition which yields minimum clutter r a d is derived. PRI-staggered postdop ler meets the condition and rovides both ezcellent per6rmance with few degrees orfreedom and the abili t y to maintain low adapted doppler sadelobes. Adjacent bin post-doppler re uires more de rees of freedom when low doppler sidelo%es are desire!","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125628763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572541
J. Praschifka
Spread clutter is a phenomenon affecting over-thehorizon radars whereby the Doppler spectrum in the vicinity of zero Hertz becomes corrupted by clutter returns, thus obscuring low velocity target signals. The suppression of spread clutter using adaptive noise cancelling techniques is analysed and the consequences for detection and tracking performance are discussed. The analysis is carried out using data from the Australian Jindalee over-the-horizon radar at Alice Springs.
{"title":"Investigation of Spread Clutter Mitigation for Oth Radar Using an Adaptive Noise Canceller","authors":"J. Praschifka","doi":"10.1109/SSAP.1994.572541","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572541","url":null,"abstract":"Spread clutter is a phenomenon affecting over-thehorizon radars whereby the Doppler spectrum in the vicinity of zero Hertz becomes corrupted by clutter returns, thus obscuring low velocity target signals. The suppression of spread clutter using adaptive noise cancelling techniques is analysed and the consequences for detection and tracking performance are discussed. The analysis is carried out using data from the Australian Jindalee over-the-horizon radar at Alice Springs.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126307794","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572483
J. M. Salavedra, E. Masgrau, A. Moreno, J. Estarellas
We study some speech enhancement algorithms based on the iterative Wiener filtering method due to Lim-Oppenheim [2], where the AR spectral estimation of the speech is carried out using a 2nd-order analysis. But in our algorithms we consider an AR estimation by means of cumulant analysis. This work extends some preceding papers due to the authors, providing a different frame length where AR estimation is done. Information of previous speech frames is used to initiate speech AR modelling of the current frame. Two parameters are introduced to dessign Wiener filter at first iteration of this iterative algorithm. These parameters are the Interframe Factor IF and the Previous Frame Iteration PFI. They allow a very important noise suppression after processing only fxst iteration of this algorithm, without any appreciable increase of distortion.
{"title":"Variable Frame Length Of A Higher Order Speech AR Estimation In A Speech Enhancement System","authors":"J. M. Salavedra, E. Masgrau, A. Moreno, J. Estarellas","doi":"10.1109/SSAP.1994.572483","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572483","url":null,"abstract":"We study some speech enhancement algorithms based on the iterative Wiener filtering method due to Lim-Oppenheim [2], where the AR spectral estimation of the speech is carried out using a 2nd-order analysis. But in our algorithms we consider an AR estimation by means of cumulant analysis. This work extends some preceding papers due to the authors, providing a different frame length where AR estimation is done. Information of previous speech frames is used to initiate speech AR modelling of the current frame. Two parameters are introduced to dessign Wiener filter at first iteration of this iterative algorithm. These parameters are the Interframe Factor IF and the Previous Frame Iteration PFI. They allow a very important noise suppression after processing only fxst iteration of this algorithm, without any appreciable increase of distortion.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134434928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572423
Jeng-Kuang Hwang, Jiunn-Horng Denq
The maximum likelihood parameter estimation of multiple damped sinusoids in noise is considered in this paper. Since the damped signal decays exponentially with time and each signal has two parameters to estimate, the ML criterion is very Mcu l t to optimize. In computing the MLE, it is noted that the convergence performance of the iterative algorithm is highly sensitive to the initial point. Thus we resort to a Newton-type ML algorithm equipped with an improved initialization scheme, which comkts of a robust state-space method followed by a reibing alternating " b a t i o n (AM) procedure. Performance simutation shows that the overall ML algorithm can achieve the CR bound with a lower threshold SNR than other existing methods. lies on how to optimize the highly nonlinear and multidimensional ML Criterion [3-51. As is well known, a key to the global convergence of the ML algorithm is the determination of the initial point. In this paper, we present a two-step initialization scheme for finding a more stable initial point. The first step is a polynomialbased state space method that can resuit in stable estimates of the damping fixtors, and the second step is a rething alternating " b a t i o n (AM) methd used to find more accurate frequency estimates [4]. Once the initialization is completed, Newton-type iterations similar to that in [5] are perfiormed in the main loop to optimize the ML criterion. In the following sections, we will present the problem formulation and the overall ML algorithm. Then its superior performance, as compared to other methods, is confirmed by computer simulations.
研究了噪声条件下多阻尼正弦波的最大似然参数估计问题。由于阻尼信号随时间呈指数衰减，并且每个信号都有两个参数需要估计，因此ML准则非常容易优化。在计算MLE时，注意到迭代算法的收敛性能对初始点高度敏感。因此，我们采用牛顿型机器学习算法，该算法配备了改进的初始化方案，该方案包括鲁棒状态空间方法，然后是控制交替的“b - a - i - o - n (AM)”过程。性能仿真表明，整体ML算法能够以较低的信噪比实现CR边界。在于如何优化高度非线性和多维的ML准则[3-51]。众所周知，ML算法全局收敛的关键是初始点的确定。在本文中，我们提出了一个两步初始化方案来寻找一个更稳定的初始点。第一步是基于多项式的状态空间方法，该方法可以得到阻尼固定器的稳定估计，第二步是一种交替的“b - a - i - o - n (AM)”方法，用于找到更准确的频率估计[4]。初始化完成后，在主循环中执行类似于[5]的牛顿型迭代来优化ML标准。在接下来的章节中，我们将介绍问题的表述和整个ML算法。并通过计算机仿真验证了该方法的优越性。
{"title":"Maximum Likelihood Estimation of Multiple Damped Sinusoids by Using Newton's Iterations and Improved Initialization","authors":"Jeng-Kuang Hwang, Jiunn-Horng Denq","doi":"10.1109/SSAP.1994.572423","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572423","url":null,"abstract":"The maximum likelihood parameter estimation of multiple damped sinusoids in noise is considered in this paper. Since the damped signal decays exponentially with time and each signal has two parameters to estimate, the ML criterion is very Mcu l t to optimize. In computing the MLE, it is noted that the convergence performance of the iterative algorithm is highly sensitive to the initial point. Thus we resort to a Newton-type ML algorithm equipped with an improved initialization scheme, which comkts of a robust state-space method followed by a reibing alternating \" b a t i o n (AM) procedure. Performance simutation shows that the overall ML algorithm can achieve the CR bound with a lower threshold SNR than other existing methods. lies on how to optimize the highly nonlinear and multidimensional ML Criterion [3-51. As is well known, a key to the global convergence of the ML algorithm is the determination of the initial point. In this paper, we present a two-step initialization scheme for finding a more stable initial point. The first step is a polynomialbased state space method that can resuit in stable estimates of the damping fixtors, and the second step is a rething alternating \" b a t i o n (AM) methd used to find more accurate frequency estimates [4]. Once the initialization is completed, Newton-type iterations similar to that in [5] are perfiormed in the main loop to optimize the ML criterion. In the following sections, we will present the problem formulation and the overall ML algorithm. Then its superior performance, as compared to other methods, is confirmed by computer simulations.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127167010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572494
S.I. Shah, L. Chaparro, A. El-Jaroudi
2. EVOLUTIONARY MAXIMUM ENTROPY ESTIMATION Using maximum entropy spectral analysis and the theThe Wold-Cramer representation [4] of a non-stationary by considering it the output of a linear timevarying system (LTV) with white noise as input: ory of the Wold-Cramer evolutionary spectrum we develop signal is the evolutionary maximum entropy @ME) estimator for non-stationary signals. The EME estimation reduces to the fitting of a time-varying autoregressive model to the Fourier coefficients of the evolutionary spectrum. The model parameters are efficientlv found bv means of the Levinson alH(n, w)ejwndZ(w) (1) gorithm. Just as in the stationary case, the EME estimator provides very good frequency resolution and can be used to obtain autoregressive models. In this paper, we present the EME estimator and discuss some of its properties. Our aim is to show that the EME estimator has analogous properties to the classical ME estimator for stationary signals.
{"title":"Properties of the Evolutionary Maximum Entropy Spectral Estimator","authors":"S.I. Shah, L. Chaparro, A. El-Jaroudi","doi":"10.1109/SSAP.1994.572494","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572494","url":null,"abstract":"2. EVOLUTIONARY MAXIMUM ENTROPY ESTIMATION Using maximum entropy spectral analysis and the theThe Wold-Cramer representation [4] of a non-stationary by considering it the output of a linear timevarying system (LTV) with white noise as input: ory of the Wold-Cramer evolutionary spectrum we develop signal is the evolutionary maximum entropy @ME) estimator for non-stationary signals. The EME estimation reduces to the fitting of a time-varying autoregressive model to the Fourier coefficients of the evolutionary spectrum. The model parameters are efficientlv found bv means of the Levinson alH(n, w)ejwndZ(w) (1) gorithm. Just as in the stationary case, the EME estimator provides very good frequency resolution and can be used to obtain autoregressive models. In this paper, we present the EME estimator and discuss some of its properties. Our aim is to show that the EME estimator has analogous properties to the classical ME estimator for stationary signals.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127498361","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1994-06-26DOI: 10.1109/SSAP.1994.572531
Fengzhen Wang, T. Lo, J. Litva, É. Bossé
This paper presents the multisensor data fusion for airborne target classification with artificial neural network. A feature set, which possesses the dominant characteristics of targets and has a certain dynamic range, is chosen. The entire system consists of identification nets (IN) and classification net (CN). Each identification network is used to extract a particular feature of the target, then the outputs of identification networks are fused by classification network, in which the neural network acts as a decision making processor. In the paper, multilayer perceptrons neural networks trained by back-propagation (BP) rule are discussed. In order to speed up the training or decrease the number of epoch in learning process, both momentum and adaptive learning rate methods are used. The simulation results show that the technique of automatic target classification using neural networks can achieve robust decision performance.
{"title":"Multisensor Automatic Target Classification with Neural Networks","authors":"Fengzhen Wang, T. Lo, J. Litva, É. Bossé","doi":"10.1109/SSAP.1994.572531","DOIUrl":"https://doi.org/10.1109/SSAP.1994.572531","url":null,"abstract":"This paper presents the multisensor data fusion for airborne target classification with artificial neural network. A feature set, which possesses the dominant characteristics of targets and has a certain dynamic range, is chosen. The entire system consists of identification nets (IN) and classification net (CN). Each identification network is used to extract a particular feature of the target, then the outputs of identification networks are fused by classification network, in which the neural network acts as a decision making processor. In the paper, multilayer perceptrons neural networks trained by back-propagation (BP) rule are discussed. In order to speed up the training or decrease the number of epoch in learning process, both momentum and adaptive learning rate methods are used. The simulation results show that the technique of automatic target classification using neural networks can achieve robust decision performance.","PeriodicalId":151571,"journal":{"name":"IEEE Seventh SP Workshop on Statistical Signal and Array Processing","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127582586","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：