Reduced complexity turbo equalization using a dynamic Bayesian network

Abstract

It is proposed that a dynamic Bayesian network (DBN) is used to perform turbo equalization in a system transmitting information over a Rayleigh fading multipath channel. The DBN turbo equalizer (DBN-TE) is modeled on a single directed acyclic graph by relaxing the Markov assumption and allowing weak connections to past and future states. Its complexity is exponential in encoder constraint length and approximately linear in the channel memory length. Results show that the performance of the DBN-TE closely matches that of a traditional turbo equalizer that uses a maximum a posteriori equalizer and decoder pair. The DBN-TE achieves full convergence and near-optimal performance after small number of iterations.

Applying Bayesian networks and belief propagation to error correction coding

“…The bn was also deployed as a turbo-equaliser to mitigate the effects of multi-path propagation. The graph produced near optimal results if the majority of the power is contained in the primary tap [5]. We conclude that a bn can be used successfully as a universal channel decoding method, provided a graph can be defined to describe the connections within the code.…”

“…For this reason only the most significant taps should be included [1]. Although the inclusion of a multi-path channel introduces many short loops to the network, Myburgh et al [5] showed that the bp algorithm will converge provided the primary tap is sufficiently dominant. By applying a minimum phase pre-filter to the received channel observations it is possible to enforce this constraint.…”

A low complexity Hopfield neural network turbo equalizer

“…This vision has successfully been implemented and demonstrated by the authors in [14] using a dynamic Bayesian network (DBN) as basis. In this paper, however, we show that using the Hopfield neural network (HNN) [15] as the underlying structure also works well, and has a number of advantages as discussed in [16].…”