Trellis-Based Conflict Resolution for Bidirectional Decision-Feedback Equalization

In this paper, we propose a new approach to bidirectional decision feedback equalization for channels with intersymbol interference (ISI). The decision feedback equalizer (DFE) is widely used in practice for ISI mitigation due to its simple implementation. However, the DFE suffers from error propagation that causes performance degradation. Bidirectional equalization of the received signal can be used to enhance the performance of the DFE. Bidirectional DFE consists of a traditional DFE in parallel with another DFE operating on the time-reversed sequence of the received signals. The results of the forward and reverse equalization can be combined with some arbitration for those sections of the message for which the equalizers produce contradictory decisions. In this paper, we propose a trellis-based conflict resolution (TBCR) algorithm for resolving conflict events between the outputs of the two DFEs. We show that the TBCR algorithm provides better performance and lower complexity compared to other previously proposed arbitration algorithms. We also propose a novel way to calculate the filter coefficients for the reverse DFE from the coefficients of the forward DFE, thus reducing the complexity of computing the reverse DFE´s coefficients.