Optimum lag and subset selection for a radial basis function equaliser

Abstract

This paper examines the application of the radial basis function (RBF) network to the modelling of the Bayesian equaliser. In particular, we study the effects of delay order d on decision boundary and attainable bit error rate (BFR) performance. To determine the optimum delay parameter for minimum BER performance, a simple BER estimator is proposed. The implementation complexity of the RBF network grows exponentially with respect to the number of input nodes. As such, the full implementation of the RBF network to realise the Bayesian solution may not be feasible. To reduce some of the implementation complexity, we propose an algorithm to perform subset model selection. Our results indicate that it is possible to reduce model size without, significant degradation in BER performance. Indexing Term: Bayesian equaliser, neural networks, RBF network, RER.

Asymptotic Bayesian decision feedback equalizer using a set of hyperplanes

“…In a previous work [23], the Bayesian solution is approximated by only using the set of dominant state pairs, which determine the asymptotic Bayesian decision boundary, in the computation of the Bayesian decision variable (16). In this study, we consider using the multiple-hyperplane detector structure of Kim and Moon [19], [20] (see Fig.…”

Determining the optimal decision delay parameter for a linear equalizer

“…It is well-known that the choice of equalizer decision delay parameter critically determines achievable bit error rate (BER) performance [3,4] . We present a simple and effective method for determining an optimal decision delay parameter that results in the best bit error rate performance for a linear equalizer.…”

Design of the optimal separating hyperplane for the decision feedback equalizer using support vector machines

“…Each of these hyperplanes is defined by a pair of dominant states in Ê ´¦½µ [8]. The following algorithm can be used to select these pairs of the dominant states that define the set of the asymptotic hyperplanes [9]:…”