文献类型：会议论文

英文题名：A neural network for parameter estimation of the exponentially damped sinusoids

作者：Xiao, Xiuchun[1,2]; Lai, Jian-Huang[2]; Wang, Chang-Dong[3]

机构：[1] College of Information, Guangdong Ocean University, China; [2] School of Information Science and Technology, Sun Yat-sen University, China; [3] School of Mobile Information Engineering, Sun Yat-sen University, China

会议论文集：Intelligence Science and Big Data Engineering - 4th International Conference, IScIDE 2013, Revised Selected Papers

会议日期：July 31, 2013 - August 2, 2013

会议地点：Beijing, China

语种：英文

外文关键词：Neural networks - Fault tolerance - Iterative methods

外文摘要：The problem of estimating the parameters of exponentially damped sinusoids (EDSs) has received very much attention in many fields. Strictly following the mathematic formulation of EDSs, we construct a specific neural network, termed EDSNN. In order to train EDSNN, a modified Levenberg-Marquardt iterative algorithm is derived. Profiting from good performance in fault tolerance of neural network, the proposed algorithm can be expected to possess a good performance in resistance to noise to some extent. Computer simulations have been conducted to apply this method to some EDSs signal models. The results substantiate the proposed EDSNN can obtain a higher precision for the parameters of the EDS component than the state-of-the-art algorithm. ? 2013 Springer-Verlag Berlin Heidelberg.