The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Pour les systèmes de communication massifs à entrées multiples et sorties multiples (MIMO), de simples détecteurs linéaires tels que le forçage zéro (ZF) et l'erreur quadratique moyenne minimale (MMSE) peuvent atteindre des performances de détection presque optimales avec une complexité de calcul réduite. Cependant, de tels détecteurs linéaires impliquent toujours une inversion de matrice compliquée, qui souffrira d'une surcharge de calcul élevée dans la mise en œuvre pratique. En raison du traitement parallèle massif et de l’efficacité de la mise en œuvre matérielle, le réseau neuronal est devenu une approche prometteuse du traitement du signal pour les futures communications sans fil. Dans cet article, nous proposons d'abord un réseau de neurones efficace pour calculer les pseudo-inverses pour tout type de matrices, basé sur la méthode améliorée de Newton, appelée PINN. Grâce à une analyse et une dérivation détaillées, les détecteurs massifs linéaires MIMO sont cartographiés sur des PINN, ce qui permet de tirer pleinement parti des résultats de la recherche sur les réseaux neuronaux, tant au niveau des algorithmes que du matériel. En outre, une méthode quasi-Newton améliorée de Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) à mémoire limitée est étudiée en tant qu'algorithme d'apprentissage des PINN afin d'obtenir un meilleur compromis performance/complexité. Les résultats de simulation valident enfin l'efficacité du schéma proposé.
Lin LI
Qinghai Normal University
Jianhao HU
University of Electronic Science and Technology of China
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Lin LI, Jianhao HU, "An Efficient Mapping Scheme on Neural Networks for Linear Massive MIMO Detection" in IEICE TRANSACTIONS on Fundamentals,
vol. E106-A, no. 11, pp. 1416-1423, November 2023, doi: 10.1587/transfun.2022EAP1132.
Abstract: For massive multiple-input multiple-output (MIMO) communication systems, simple linear detectors such as zero forcing (ZF) and minimum mean square error (MMSE) can achieve near-optimal detection performance with reduced computational complexity. However, such linear detectors always involve complicated matrix inversion, which will suffer from high computational overhead in the practical implementation. Due to the massive parallel-processing and efficient hardware-implementation nature, the neural network has become a promising approach to signal processing for the future wireless communications. In this paper, we first propose an efficient neural network to calculate the pseudo-inverses for any type of matrices based on the improved Newton's method, termed as the PINN. Through detailed analysis and derivation, the linear massive MIMO detectors are mapped on PINNs, which can take full advantage of the research achievements of neural networks in both algorithms and hardwares. Furthermore, an improved limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) quasi-Newton method is studied as the learning algorithm of PINNs to achieve a better performance/complexity trade-off. Simulation results finally validate the efficiency of the proposed scheme.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2022EAP1132/_p
Copier
@ARTICLE{e106-a_11_1416,
author={Lin LI, Jianhao HU, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={An Efficient Mapping Scheme on Neural Networks for Linear Massive MIMO Detection},
year={2023},
volume={E106-A},
number={11},
pages={1416-1423},
abstract={For massive multiple-input multiple-output (MIMO) communication systems, simple linear detectors such as zero forcing (ZF) and minimum mean square error (MMSE) can achieve near-optimal detection performance with reduced computational complexity. However, such linear detectors always involve complicated matrix inversion, which will suffer from high computational overhead in the practical implementation. Due to the massive parallel-processing and efficient hardware-implementation nature, the neural network has become a promising approach to signal processing for the future wireless communications. In this paper, we first propose an efficient neural network to calculate the pseudo-inverses for any type of matrices based on the improved Newton's method, termed as the PINN. Through detailed analysis and derivation, the linear massive MIMO detectors are mapped on PINNs, which can take full advantage of the research achievements of neural networks in both algorithms and hardwares. Furthermore, an improved limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) quasi-Newton method is studied as the learning algorithm of PINNs to achieve a better performance/complexity trade-off. Simulation results finally validate the efficiency of the proposed scheme.},
keywords={},
doi={10.1587/transfun.2022EAP1132},
ISSN={1745-1337},
month={November},}
Copier
TY - JOUR
TI - An Efficient Mapping Scheme on Neural Networks for Linear Massive MIMO Detection
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1416
EP - 1423
AU - Lin LI
AU - Jianhao HU
PY - 2023
DO - 10.1587/transfun.2022EAP1132
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E106-A
IS - 11
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - November 2023
AB - For massive multiple-input multiple-output (MIMO) communication systems, simple linear detectors such as zero forcing (ZF) and minimum mean square error (MMSE) can achieve near-optimal detection performance with reduced computational complexity. However, such linear detectors always involve complicated matrix inversion, which will suffer from high computational overhead in the practical implementation. Due to the massive parallel-processing and efficient hardware-implementation nature, the neural network has become a promising approach to signal processing for the future wireless communications. In this paper, we first propose an efficient neural network to calculate the pseudo-inverses for any type of matrices based on the improved Newton's method, termed as the PINN. Through detailed analysis and derivation, the linear massive MIMO detectors are mapped on PINNs, which can take full advantage of the research achievements of neural networks in both algorithms and hardwares. Furthermore, an improved limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) quasi-Newton method is studied as the learning algorithm of PINNs to achieve a better performance/complexity trade-off. Simulation results finally validate the efficiency of the proposed scheme.
ER -