Channel Equalization for OFDM System Based on the BP Neural Network

OFDM modulation combines the advantages of high achievable rates and relatively easy implementation. In this paper, we show how to perform blind maximum-likelihood data recovery in OFDM transmission from the output symbol and cyclic pre x. Our approach relies on decomposing the OFDM channel into two subchannels (cyclic and linear) that share the same input and are characterized by the same channel parameters. This fact enables us to estimate the channel parameters from one subchannel and substitute the estimate into the other, thus obtaining a nonlinear relationship involving the input and output data only. The relationship does not show any channel dependence whatsoever and can be exhaustively searched for the maximum likelihood estimate of the input. This shows that OFDM systems are completely identi able using output data only, irrespective of the channel zeros, as long as the channel delay spread is less than the length of the cyclic pre x.

show abstract

“…This comes contrary to the common belief that OFDM using CP cannot be equalized for channels with zeros on the FFT grid[1] and[6] …”

contrasting

confidence: 75%

Blind maximum-likelihood data recovery in OFDM

Al-Naffouri

Quadeer

2008

2008 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

show abstract

“…Only lately, machine learning algorithms have been in the spotlight of the optical communications community 40 . They are being considered for optical network monitoring and optimization 41,42,43 , optical header recognition 44 and mitigation of transmission effects 45,46,47,48,49,50,51,52,53,54 . Still, the drawback of applying these standard tools in ultrafast systems is that they are computationally expensive and still far away from reaching real-time processing at telecom data rates.In the present work, we provide a first validation that neuro-inspired information processing based on photonic implementations can address critical issues in the field of signal processing for high-speed communications.Specifically, we demonstrate that techniques like ELM and RC can offer solutions to data recovery of distorted signals from extended fibre transmission.…”

mentioning

confidence: 99%

Photonic machine learning implementation for signal recovery in optical communications

Argyris

Bueno

Fischer

2018

Sci Rep

141

View full text Add to dashboard Cite

Machine learning techniques have proven very efficient in assorted classification tasks. Nevertheless, processing time-dependent high-speed signals can turn into an extremely challenging task, especially when these signals have been nonlinearly distorted. Recently, analogue hardware concepts using nonlinear transient responses have been gaining significant interest for fast information processing. Here, we introduce a simplified photonic reservoir computing scheme for data classification of severely distorted optical communication signals after extended fibre transmission. To this end, we convert the direct bit detection process into a pattern recognition problem. Using an experimental implementation of our photonic reservoir computer, we demonstrate an improvement in bit-error-rate by two orders of magnitude, compared to directly classifying the transmitted signal. This improvement corresponds to an extension of the communication range by over 75%. While we do not yet reach full real-time post-processing at telecom rates, we discuss how future designs might close the gap.

show abstract

“…Although the traditional BP algorithm works in real domain, signals to be equalized in radio frequency OFDM systems are complex. Therefore, in radio frequency OFDM equalization studies, the algorithm has been extended to the complex domain [19,20].…”

Section: A Back-propagation Neural Equalizermentioning

confidence: 99%

“…This algorithm works on a multi-layer perceptron which is a kind of neural networks with several hidden layers. BP algorithm achieves to the global minimization by steepest descent method [19]. Although the traditional BP algorithm works in real domain, signals to be equalized in radio frequency OFDM systems are complex.…”

Section: A Back-propagation Neural Equalizermentioning

confidence: 99%

Performance of back-propagation and self organizing map neural equalizers for asymmetrically clipped optical OFDM

Javidi

Khoshbin

2013

2013 21st Iranian Conference on Electrical Engineering (ICEE)

View full text Add to dashboard Cite

This paper evaluates the performance of backpropagation (BP) and self organizing map (SOM) equalizer for indoor asymmetrically clipped optical OFDM (ACO-OFDM) wireless systems. Although in OFDM, parallel transmission and cyclic prefix addition improve the communication efficiency, there is still performance degradation due to inter-symbol interference (ISI) in dispersive channels. Simulation results indicate that BP multilayer perceptron and SOM neural equalizers enhance the bit error rate performance of ACO-OFDM systems in diffused channels for high signal to noise ratio. Moreover, proposed BP and SOM equalization require only 0.05 and 0.005 percent of ACO-OFDM symbols for training in a sec respectively while in single-tap equalization channel state information is necessary at the receiver.

show abstract

Channel Equalization for OFDM System Based on the BP Neural Network

Cited by 21 publications

References 6 publications

Blind maximum-likelihood data recovery in OFDM

Blind maximum-likelihood data recovery in OFDM

Photonic machine learning implementation for signal recovery in optical communications

Performance of back-propagation and self organizing map neural equalizers for asymmetrically clipped optical OFDM

Contact Info

Product

Resources

About