Premium
Asymptotic optimal detection for MIMO communication systems employing tree search with incremental channel partition preprocessing
Author(s) -
Radji Djelili,
Leib Harry
Publication year - 2013
Publication title -
transactions on emerging telecommunications technologies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.366
H-Index - 47
ISSN - 2161-3915
DOI - 10.1002/ett.2538
Subject(s) - mimo , computational complexity theory , algorithm , partition (number theory) , tree traversal , preprocessor , tree (set theory) , channel (broadcasting) , computer science , mathematics , variable (mathematics) , telecommunications , mathematical analysis , combinatorics , artificial intelligence
The high complexity of optimal detection for spatial multplexing multiple‐input multiple‐output systems motivates the need for more practical alternatives. Among many suboptimal schemes reported in the literature, very few can be proven to provide close to optimal performance with low fixed complexity. The recently introduced Selection based Minimum Mean Square Error Ordered Successive Interference Cancellation (Sel‐MMSE‐OSIC) algorithm is one such scheme that employs list‐based detection. Simulations results showed that its performance is nearly indistinguishable from optimal at almost all signal‐to‐noise ratio (SNR) levels. In this paper, we propose an improved asymptotically optimal fixed‐complexity algorithm that provides substantial complexity reductions over Sel‐MMSE‐OSIC with similar error rate performance. This scheme is based on simplified channel partition and efficient tree‐based list detection. To achieve further reductions in complexity for large constellation sizes, a variable complexity version of this scheme is proposed. The resulting algorithm is a variable complexity scheme that operates on a very small subset of candidates and employs an improved channel partition preprocessing that not only reduces complexity but also guarantees high SNR optimality over space uncorrelated channels. Simulations results confirm that the proposed scheme provides significant complexity reductions over conventional variable complexity detection schemes. Copyright © 2012 John Wiley & Sons, Ltd.