Robust iterative tree-pruning detection and LDPCC decoding

A novel suboptimal low-complexity equalization and turbo-iterative decoding scheme is proposed in this paper. The scheme is developed for multiple transmit- and multiple receive-antenna systems operating over severe frequency-selective fading intersymbol interference channels. The signal-processing complexity may be of a concern for such systems. The complexity of a full-search equalization grows in a power-law manner O(M^NtL), where M denotes for M-ary channel symbols, N_t the number of transmit-antennas, and L the number of delay channel taps. A low-complexity solution can be obtained by pruning an equalizer tree. The two main operations include a sphere list detection and a threshold-based tree-search. In the operation of extracting extrinsic messages from the pruned tree, a set of explored paths with different survival lengths poses a fairness problem: a longer-lived path naturally builds a larger discrepancy-metric than a shorter lived path does. A novel survival-length compensation-rule is devised so that all explored paths with different survival lengths are utilized fairly in generating the output message. Simulation results are obtained for multi-input and multi-output systems equipped with four transmit and four receive antennas. They indicate the performance of the receiver is very robust.