On Complex LLL Algorithm for Integer Forcing Linear Receivers

TL;DR

This paper introduces a low-complexity complex LLL algorithm for selecting integer coefficients in integer-forcing MIMO receivers, demonstrating improved ergodic rate and BER performance over traditional linear receivers.

Contribution

It proposes a novel low-complexity complex LLL-based method for integer coefficient selection in IF receivers, enhancing practicality and performance.

Findings

Outperforms MMSE and ZF in BER

Achieves comparable ergodic rate to more complex methods

Offers a favorable complexity-performance trade-off

Abstract

Integer-forcing (IF) linear receiver has been recently introduced for multiple-input multiple-output (MIMO) fading channels. The receiver has to compute an integer linear combination of the symbols as a part of the decoding process. In particular, the integer coefficients have to be chosen based on the channel realizations, and the choice of such coefficients is known to determine the receiver performance. The original known solution of finding these integers was based on exhaustive search. A practical algorithm based on Hermite-Korkine-Zolotareff (HKZ) and Minkowski lattice reduction algorithms was also proposed recently. In this paper, we propose a low-complexity method based on complex LLL algorithm to obtain the integer coefficients for the IF receiver. For the 2 X 2 MIMO channel, we study the effectiveness of the proposed method in terms of the ergodic rate. We also compare the bit error rate (BER) of our approach with that of other linear receivers, and show that the suggested algorithm outperforms the minimum mean square estimator (MMSE) and zero-forcing (ZF) linear receivers, but trades-off error performance for complexity in comparison with the IF receiver based on exhaustive search or on HKZ and Minkowski lattice reduction algorithms.