Theoretical derivation of minimum mean square error of RBF based equalizer
Introduction
In digital communication systems, time dispersion caused by the non-ideal channel frequency response characteristics or multi-path transmission creates overlapping of the received symbols due to intersymbol interference (ISI). The conventional way to combat ISI is to include an equalizer in the receiver. The most widely known equalizer is an adaptive linear transversal equalizer, in which the output signal is compared to the expected signal and the tap coefficients are updated in accordance with the error between the desired signal and actual filter output. For more than a decade, there has been much attention given to applying neural networks to the digital communication areas, including channel equalization problems [1]. Multi-layer perceptrons (MLP) equalizer [2], [3], [4] is able to equalize non-minimum phase channels without the introduction of any time delay; and it is less susceptible than a linear equalizer to the effects of high levels of additive noise. However, the network architecture and training algorithm of the MLP equalizer is much more complex than the linear equalizer. Also, the RBF network has received a great deal of attention by many researchers because of its structural simplicity and more efficient learning [5], [6], [7], [8], [9], [10], [11], [12]. In [5], Chen et al. applied RBF network to channel equalization problem to get the optimal Bayesian solution. Also, the application of RBF network to the modeling of Bayesian equalizer was discussed [13]. In this paper, the effect of delay order on making decision boundary and optimum delay for minimum bit error rate performance was studied. Although the RBF based equalizers provide great advantages over MLP based equalizer, they still require considerable number of RBF centers when channel order goes high, which finally leads to computational burden in training. To solve this problem, some researchers developed techniques for either reducing the number of centers or selecting them wisely for training [6], [8], [10], [12]. In a paper by Kumar et al. [12], the minimal RBF network [11] was applied to channel equalization. It says that channel equalizer with minimal RBF network provide efficient solution without requiring much complexities, such as channel order estimation and fixing of the number of channel states.
Although many of studies, mentioned above, claims that RBF based equalizers are superior to conventional linear equalizer due to both RBF network's structural linearity (or simplicity) and efficient training, none of them tried to compare those two from the theoretical minimum mean square error (MSE) point of view. The basic idea of comparing these two equalizers comes from the fact that once input domain of RBF network is transformed to another domain through Gaussian basis functions, the relationship between the hidden and output layers in the RBF equalizer is also linear. In this paper, the theoretical minimum MSE for various channels for both RBF and the linear equalizers were evaluated and compared. Also, the sensitivity of minimum MSE to equalizer order, channel delay, and RBF center spreads was analyzed.
Section snippets
Background of RBF equalizer
A digital information source is to transmit independent and equi-probable binary symbols, designated as , that are either or . Symbols are transmitted in modulated form through a channel which is dispersive, causing the symbols to spread in time and produce ISI. Dispersion in the digital channel may be represented by the transfer functionwhere p is the channel order and d the channel delay. That is, the channel output value may contain interference from
Derivation of minimum MSEs of RBF equalizer
The main objective of this paper is to derive the theoretical minimum MSE of RBF based equalizer and compare it with that of the linear equalizer (finite transversal filter). The basic idea of comparing these two equalizers comes from the fact that the relationship between the hidden and output layers in the RBF equalizer is also linear. To do this, first the general theory of minimum MSE for linear equalizer was reviewed [15]. The theoretical minimum MSE of linear filter is as follows:
Simulation studies by theoretical approach
The following channel model is selected as an example for calculating and comparing the for both RBF and linear equalizer [5]The equalizer order, q, is assumed to be 1. As mentioned earlier, the transmitted sequences are assumed to consist of independent and equi-probable binary symbols, designated as , that are either or . Then the maximum number of RBF centers, M is equal to 8 . To determine , the noise-free channel output, and
Conclusion
Traditional adaptive algorithms for channel equalizers are based on the criterion of minimizing the MSE between the desired filter output and the actual filter output. In this paper, the theoretical minimum MSE of a RBF equalizer was evaluated and compared with that of a linear equalizer. The procedure of computing theoretical minimum MSE of RBF equalizer was derived using the same concepts of finding minimum MSE for liner equalizer, based on the fact that the relationship between the hidden
References (16)
- M. Ibnkahla, Applications of neural networks to digital communications—a survey, Signal Processing (1999)...
- G.J. Gibson, S. Siu, C.F.N. Cowan, Application of multilayer perceptrons as adaptive channel equalizers, in:...
- et al.
Adaptive decision feedback equalization for digital channels using multilayer neural networks
IEEE J. Selected Areas Comm.
(1995) - et al.
The use of neural nets to combine equalization with decoding for severe intersymbol interference channels
IEEE Trans. Neural Networks
(1994) - et al.
A clustering technique for digital communication channel equalization using radial basis function networks
IEEE Trans. Neural Networks
(1993) - et al.
Reduced complexity implementation of Bayesian equalizer using local RBF networks for channel equalization problem
Electron. Lett.
(1996) Applying radial basis functions
IEEE Signal Proc. Mag.
(1996)- et al.
Design for centers of RBF neural networks for fast time-varying channel equalization
Electron. Lett.
(1996)
Cited by (22)
An adaptive decision feedback equalizer based on the combination of the FIR and FLNN
2011, Digital Signal Processing: A Review JournalPipelined functional link artificial recurrent neural network with the decision feedback structure for nonlinear channel equalization
2011, Information SciencesCitation Excerpt :It has been applied to channel equalization to overcome various distortions. However, more RBF centers are required when the channel order goes high, which finally results in heavy computational burden [4,15]. To overcome the shortcomings of the MLP, RNN and RBF, a single layer polynomial perceptron network without any hidden layer has been utilized for channel equalization in which the original input pattern is expanded by the truncated Volterra series.
Functional link neural network cascaded with Chebyshev orthogonal polynomial for nonlinear channel equalization
2008, Signal ProcessingCitation Excerpt :The adaptive algorithms are more easily used to train the network and have lower complexity because of the absence of any hidden layers. Applications of FLNN for channel equalization and nonlinear system identification have been reported in [8,10,11], the performance of the FLNN-based equalizer is superior to other neural network structures such as MLP for both linear and nonlinear channel models, and the main advantages of FLNN are the further reduced computation cost. Though the FLNN have some advantages, such as simpler structure, faster convergence and lower computational complexity, the nonlinear approximation capacity is limited due to only one nonlinear function tanh(.).
A Novel Cuckoo Search Optimized RBF Trained ANN in a Nonlinear Channel Equalization
2023, Lecture Notes in Networks and SystemsImproving the modeling and forecasting of fuel selling price using the radial basis function technique: A case study
2019, Journal of Algorithms and Computational Technology