Efficient computation of the realizable MIMO DFE

Realizable DFEs are DFEs with stable and causal IIR filters and finite decision delay. Computational complexity of current algorithms to compute them usually grows cubically with the decision delay. In this paper, we show how complexity can be reduced to quadratic. We compare two approaches, the so-called polynomial approach and a novel state-space approach using inner-outer factorization. In both cases finite linear equation systems with structure lie at the heart of the realizable DFE. Displacement structure theory allows to solve them efficiently.