• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.11
• /
• pp.917-924
• /
• 2003

Recurrent neural networks have been successfully applied to communications channel equalization. Major disadvantages of gradient-based learning algorithms commonly employed to train recurrent neural networks are slow convergence rates and long training sequences required for satisfactory performance. In a high-speed communications system, fast convergence speed and short training symbols are essential. We propose decision feedback equalizers using a recurrent neural network trained with Kalman filtering algorithms. The main features of the proposed recurrent neural equalizers, utilizing extended Kalman filter (EKF) and unscented Kalman filter (UKF), are fast convergence rates and good performance using relatively short training symbols. Experimental results for two time-varying channels are presented to evaluate the performance of the proposed approaches over a conventional recurrent neural equalizer.

• Journal of Communications and Networks
• /
• v.9 no.1
• /
• pp.34-42
• /
• 2007

Orthogonal frequency division multiplexing (OFDM) combined with time division multiplexing (TDM), in this paper called OFDM/TDM, can overcome the high peak-to-average-power ratio (PAPR) problem of the conventional OFDM and improve the robustness against long time delays. In this paper, the bit error rate (BER) performance of OFDM/FDM in a frequency-selective Rayleigh fading. channel is evaluated by computer simulation. It is shown that the use of frequency-domain equalization based on minimum mean square error criterion (MMSE-FDE) can significantly improve the BER performance, compared to the conventional OFDM, by exploiting the channel frequency-selectivity while reducing the PAPR or improving the robustness against long time delays. It is also shown that the performance of OFDM/FDM designed to reduce the PAPR can bridge the conventional OFDM and single-carrier (SC) transmission by changing the design parameter.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.10
• /
• pp.2247-2253
• /
• 2010

WiBro(Wireless Broadband Internet) is the standard of high-speed portable internet based on OFDMA/TDD (Orthogonal frequency division multiple access / Time division duplexing) techniques, and the subset of consolidated version of IEEE802.16e Wireless MAN standard. In this paper, we propose performance improvements of WiBro system using the LVQ(Learning Vector Quantization) blind equalization. Proposed method used the prefiltering LVQ neural network blind equalization in the Broadband WiBro system receiver. The prefiltering LVQ neural network constellates 16QAM that is transmitter data shape and the blind equalization removes ICI(Inter Carrier Interference). To verificate the proposed method usability, the MSE(Mean Square Error) and the BER(Bit Error Rate) are simulated. The simulation results shown that is improved the performances of the proposed WiBro system using the LVQ blind equalization than the existing WiBro system.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.9
• /
• pp.1638-1645
• /
• 2008

WiBro(Wireless Broadband Internet) is the standard of high-speed portable internet based on OFDMA/TDD techniques, and the subset of consolidated version of IEEE802.16e Wireless MAN standard. In this paper, we propose the design method of WiBro system using the SOFM Blind Equalization. Proposed design method used SOFM neural network blind equalization with Bussgang algorithms in the Broadband 16 QAM WiBro system receiver. To verificate the proposed design method usability, the MSE and the BER are simulated. The simulation results shown that is improved the equalization performances of the proposed WiBro system using the SOFM Blind equalization than the existing WiBro system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.5
• /
• pp.504-511
• /
• 2016

In this paper, MIMO system based on turbo equalization techniques which LDPC codes were outer code and space time trellis codes (STTC) were employed as an inner code are studied. LDPC decoder and STTC decoder are connected through the interleaving and de-interleaving that updates each other's information repeatedly. In conventional turbo equalization of MIMO system, BCJR decoder which decodes STTC coded bits required two-bit wise decoding processing. Therefore duo-binary turbo codes are optimal for MIMO system combined with STTC codes. However a LDPC decoder requires bit unit processing, because LDPC codes can't be applied to these system. Therefore this paper proposed turbo equalization for MIMO system based on LDPC codes combined with STTC codes. By the simulation results, we confirmed performance of proposed turbo equalization model was improved about 0.6dB than that of conventional LDPC codes.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.6
• /
• pp.9-18
• /
• 1998

In this paper, an equalization technique for OFDM(orthogonal frequency division multiplexing) in a time-variant multipath fading environment is described. A loss of subchannel orthogonality due to time-varying multipath fading channels leads to interchannel interference (ICI) which increases the error floor in proportion to Doppler frequency. A simple frequency-domain equalizer which can compensate the effect of ICI caused by time variation of multipath fading channel is proposed by modifying the previous frequency-domain equalization technique with taking into account only the ICI terms significantly affecting the error performance. The effectiveness of the proposed approach is demonstrated via computer simulation by applying it to OFDM systems when the multipath fading channel is slowly time variant.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.11
• /
• pp.2116-2132
• /
• 1994

In this paper, channel equalization techniques for full-digital HDTV systems are investigated Conventional equalization methods are surveyed and several channel are modeled for computer simulation. A VS-LMS (Variable Step size Least Mean Square) algorithm using the time constant concept is proposed and its performance is compared. Several equalization techniques for HDTV systems are simulated based on various channel models, and their characteristics are analyzed. Also the equalizer using fixed-point operations is simulated and its filter structure suitable for high bit rate transmission is also studied.

• Proceedings of the KIPE Conference
• /
• 2007.07a
• /
• pp.347-349
• /
• 2007

Charge equalization scheme for HEV lithium-ion battery system is proposed in this paper, where all the primary windings with in parallel bi-directional switches are coupled in series to provide the equalizing energy from the whole battery string to the specific under charged cells. Moreover, to realize minimized size of equalization circuit employing the proposed cell balancing scheme, the optimal power rating design rule according to equalization time and SOC distribution of imbalance is proposed. A prototype of HEV lithium-ion battery system of four cells shows the outstanding charge equalization performance while maintaining greatly reduced size of cell balancing circuit.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.445-447
• /
• 2005

MCMA(modified constant modulus algorithm) accomplishes blind equalization and carrier phase recovery simultaneously. But, the error level of MCMA is not zero when the equalizer converges completely. Because the MCMA uses a special signal point instead of a original signal point. MCMA-DO(decision-directed) improves the steady-state performance but the performance of equalizer is decided by switching time between the MCMA and the DD. In this paper, according to the residual ISI(intersymbol interference) of the equalizer output, the most suitable switching time is decided automatically.

• Proceedings of the KIPE Conference
• /
• 2002.11a
• /
• pp.186-189
• /
• 2002

Real-time thermal image is obtained by thermal imaging systems measuring radiant emittence according to law of Plank's blackbody radiation. The histogram of thermal image is not uniform. The signal bands of background and target are separated and grouped in narrow bands. In such a system, contrast enhancement indispensible to distinguish target from background. In this study, plateau histogram equalization using local histogram is proposed for contrast enhancement.