軟判定値出力LRA-MIMO復号器の性能改善に関する一検討

摘要

Recently, Lattice Reduction Aided (LRA) Detection has attracted the attention of many researchers. As a Multiple-Input Multiple-Output (MIMO) decoder, it offers similar decoding accuracy as Maximum-Likelihood Detection (MLD) but with relatively lower complexity. In real systems, LRA-Detection is combined with forward error correction coding to further improve its performance. In order to obtain a higher error correction effect, soft-output LRA-Detection using the Log-Likelihood Ratio (LLR) has been studied. Conventional methods for soft-output generation in MIMO-OFDM systems are not optimal. This is because it does not consider the noise enhancement component which differs in each subcarrier. In this paper, we propose a soft-output LRA-Detection method by considering the noise enhancement component per subcarrier. We show the superiority of our proposed algorithm in IEEE 802.11ac next generation wireless LAN based systems through computer simulations.%MIMO（Multiple-Input Multiple-Output）復号法としてMLD（Maximum-Likelihood Detection）とほぼ同等の復号精度が得られ,比較的低演算量であるLRA-Detection（Lattice Reduction Aided Detection）が注目されている.このLRA-Detectionにおいては，より高い誤り訂正効果を得るために信頼性の情報であるLLR（Log Likelihood Ratio）を使用する軟判定値出力LRA-Detectionが研究されている．しかし，従来法の軟判定値出力LRA-Detectionは，OFDM（Orthogonal Frequency Division Multiplexing）無線通信システム特有のサブキャリア毎に変化する雑音強調成分が考慮されていないため，最適なLLR値ではない．本報告では，サブキャリア毎の雑音強調成分を考慮した軟判定値出力LRADetection復号の重み付け手法について提案する．また，従来法との演算量の比較，次世代無線LAN規格IEEE802.11acに準拠したシミュレータによる性能評価を行うことで，従来法に対する優位性を示す．