Low-complexity Seysen's algorithm based lattice reduction-aided MIMO detection for hardware implementations

• Published in: 2010 Conference Record of the Forty Fourth Asilomar Conference on Signals, Systems and Computers pp. 1468 - 1472
• Main Authors: Bruderer, L, Senning, C, Burg, A
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2010
• Subjects: Computational complexity; Hardware; Lattices; Measurement; MIMO; Runtime
• ISBN: 1424497221; 9781424497225
• ISSN: 1058-6393; 2576-2303
• DOI: 10.1109/ACSSC.2010.5757780

Lattice reduction-aided linear detectors for MIMO systems are a promising receiver structure for low-complexity implementations. In this paper we present a lattice reduction-aided MIMO data detection architecture based on Seysen's algorithm, where the algorithm is carried out exclusively on the Gram matrix of the channel and its inverse. Furthermore, we describe and evaluate several modification of Seysen's algorithm tailored to reduce the total complexity with respect to hardware implementation. By means of complexity-performance trade-offs we demonstrate the potential benefit of the various algorithmic modifications. First, novel schemes to identify the update steps in each iteration of Seysen's algorithm are presented. Second, we show that further complexity reduction of Seysen's algorithm can be obtained by severely constraining the update coefficients. Eventually, methods are devised that terminate Seysen's algorithm prematurely and thus result in a reduction of the average complexity.