Performance-complexity trade-offs of the 2-D iterative feedback signal detection algorithm

We present a novel iterative signal detection algorithm for two-dimensional (2-D) intersymbol-interference (ISI) channels based on a modified version of the one-dimensional (1-D) maximum a-posteriori (MAP) detection algorithm. The detector operates iteratively by combining two separate 1-D detectors acting along rows and columns using 2-D local feedback derived from previous extrinsic information. The algorithmic framework is flexible to incorporate various state and feedback configurations. By using a single-row/column detector, we can trade-off 0.4 dB of SNR performance with a fifteen times less complex realization. The performance of the algorithm is close to the maximum-likelihood (ML) estimate [17].