Decision feedback recurrent neural equalization with fast convergence rate

被引：28

作者：

Choi, J ^{[1
]}

Bouchard, M ^{[1
]}

Yeap, TH ^{[1
]}

机构：

[1] Univ Ottawa, Sch Informat Technol & Engn, Ottawa, ON K1N 6N5, Canada

来源：

IEEE TRANSACTIONS ON NEURAL NETWORKS | 2005年 / 16卷 / 03期

关键词：

channel equalization; extended Kalman filter (EKF); real-time recurrent learning (RTRL); recurrent neural network (RNN); time-varying channel;

D O I：

10.1109/TNN.2005.845142

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Real-time recurrent learning (RTRL), commonly employed for training a fully connected recurrent neural network (RNN), has a drawback of slow convergence rate. In the light of this deficiency, a decision feedback recurrent neural equalizer (DFRNE) using the RTRL requires long training sequences to achieve good performance. In this paper, extended Kalman filter (EKF) algorithms based on the RTRL for the DFRNE are presented in state-space formulation of the system, in particular for complex-valued signal processing. The main features of global EKF and decoupled EKF algorithms are fast convergence an good tracking performance. Through nonlinear channel equalization, performance of the DFRNE with the EKF algorithms is evaluated and compared with that of the DFRNE with the RTRL.

引用

页码：699 / 708

页数：10

共 24 条

[1] ADAPTIVE BAYESIAN EQUALIZER WITH DECISION-FEEDBACK [J].

CHEN, S ;

MULGREW, B ;

MCLAUGHLIN, S .

IEEE TRANSACTIONS ON SIGNAL PROCESSING, 1993, 41 (09) :2918-2927

[2] ADAPTIVE EQUALIZATION OF FINITE NONLINEAR CHANNELS USING MULTILAYER PERCEPTRONS [J].

CHEN, S ;

GIBSON, GJ ;

COWAN, CFN ;

GRANT, PM .

SIGNAL PROCESSING, 1990, 20 (02) :107-119

[3]

Coelho PHG, 2001, IEEE IJCNN, P120, DOI 10.1109/IJCNN.2001.939003

[4]

Coelho PHGA, 2001, IEEE IJCNN, P1756, DOI 10.1109/IJCNN.2001.938427

[5]

Cowan CFN, 1998, INT J ADAPT CONTROL, V12, P195, DOI 10.1002/(SICI)1099-1115(199803)12:2<195::AID-ACS487>3.0.CO

[6]

2-K

[7] Performance comparison among neural decision feedback equalizers [J].

Di Claudio, ED ;

Parisi, R ;

Orlandi, G .

IJCNN 2000: PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOL V, 2000, :361-365

[8] A signal processing framework based on dynamic neural networks with application to problems in adaptation, filtering, and classification [J].

Feldkamp, LA ;

Puskorius, GV .

PROCEEDINGS OF THE IEEE, 1998, 86 (11) :2259-2277

[9] An improved recurrent neural network for M-PAM symbol detection [J].

Hacioglu, K .

IEEE TRANSACTIONS ON NEURAL NETWORKS, 1997, 8 (03) :779-783

[10]

Haykin S., 1999, Neural Networks: A Comprehensive Foundation, V2nd ed

← 1 2 3 →