Moonsik Min

Prof. Moonsik Min

orcid

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 September 2021 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
KKKyeongbo Kong
Kyeongbo Kong
WSWoo-Jin Song
Woo-Jin Song
MMMoonsik Min
Moonsik Min
We propose a deep learning-based channel estimation, quantization, feedback, and precoding method for downlink multiuser multiple-input and multiple-output systems. In the proposed system, channel estimation and quantization for limited feedback are handled by a receiver deep neural network (DNN). Precoder selection is handled by a transmitter DNN. To emulate the traditional channel quantization, a binarization layer is adopted at each receiver DNN, and the binarization layer is also used to enable end-to-end learning. However, this can lead to inaccurate gradients, which can trap the receiver DNNs at a poor local minimum during training. To address this, we consider knowledge distillation, in which the existing DNNs are jointly trained with an auxiliary transmitter DNN. The use of an auxiliary DNN as a teacher network allows the receiver DNNs to additionally exploit lossless gradients, which is useful in avoiding a poor local minimum. For the same number of feedback bits, our DNN-based precoding scheme can achieve a higher downlink rate compared to conventional linear precoding with codebook-based limited feedback.

11 October 2016 in IEEE Transactions on Communications
Verified
Journal Article
Closed access
MMMoonsik Min
Moonsik Min
YJYo-Seb Jeon
Yo-Seb Jeon
GIGi-Hong Im
Gi-Hong Im
We consider the block diagonalization (BD) and user selection based on limited feedback in multiple antenna broadcast channels. With limited feedback, due to the imperfect channel state information at the transmitter (CSIT), BD cannot completely eliminate multiuser interference, and the throughput is correspondingly lower than that achieved with perfect CSIT. Nevertheless, the achievable multiuser diversity gain can be the same, such that limited-feedback-based BD can achieve an optimal throughput growth as the number of users increases. To show this, we first propose a channel quality indicator (CQI) for user selection. The CQI is designed to accurately estimate an achievable rate of each user and is given by an expected rate, where the expectation is solely taken over precoding matrices which cannot be known at the feedback stage. With the proper CQI, user selection can benefit from a large number of users in the system. As a result, we show that the BD can achieve an asymptotically optimal growth in throughput with the proposed CQI, based solely on a finite-rate feedback of channel information.

8 May 2018 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
MMMoonsik Min
Moonsik Min
A cellular system in which multi-antenna base stations (BSs) communicate with single-antenna users is considered. BSs are distributed according to a Poisson point process and each BS uses zero-forcing beamforming (ZFBF) to simultaneously transmit multiple data streams prepared for multiple users. Single-user transmission, which is equivalent to maximum ratio transmission, is also considered as a special case of ZFBF. Net spectral efficiency is the main performance metric used in this paper, and is analyzed theoretically to obtain the optimal number of feedback bits required to maximize it. Specifically, tight upper and lower bounds for the derivative of spectral efficiency are newly determined and the bounds are further used to estimate the optimal number of feedback bits. Consequently, a close estimate of the optimal number is proposed and the estimate accurately shows the scaling rate of the optimal number. The analytical results of this paper can provide more accurate insights into the required number of feedback bits and achievable system-level performance in interference-limited cellular networks.

20 March 2018 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
YJYo-Seb Jeon
Yo-Seb Jeon
MMMoonsik Min
Moonsik Min
Two-stage beamforming is a method that uses two types of beamformers, to reduce the feedback overhead of frequency-division-duplexing (FDD) massive multiple-input multiple-output (MIMO) systems that are spatially correlated. In this paper, we present large system analysis of two-stage beamforming when a transmitter has limited channel information from feedback and when regularized-zero-forcing (RZF) is used as a second-stage beamformer. For the feedback, we consider two different methods: One developed for when first-stage beamformer information at a receiver is available and one for when it is not. For each feedback method, we analyze both expected signal-to-interference-plus-noise ratio (SINR) and expected rate loss, and characterize their bounds as a function of the number of feedback bits. Analysis results reveal that the sum rate of two-stage beamforming is very sensitive to the regularization parameter of RZF, especially when the number of feedback bits is limited. Therefore, we also derive the optimal regularization parameter that maximizes the SINR of two-stage beamforming with limited feedback. Using simulations, we verify the analysis and quantify how the use of the optimal regularization parameter increases the sum rate of two-stage beamforming.

12 May 2022 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
TKtae-kyoung kim
tae-kyoung kim
MMMoonsik Min
Moonsik Min
YJYo-Seb Jeon
Yo-Seb Jeon
This study explores the performance of multiple-input multiple-output (MIMO) systems with one-bit analog-to-digital converters (ADCs) over Rayleigh fading channels. We particularly aim to characterize the pairwise error probability (PEP) achieved by maximum likelihood detection (MLD), which provides the optimal detection performance of the considered systems. First, the upper bound of the PEP is derived by leveraging Jensen’s inequality and the Chernoff bound. The asymptotic bounds of the PEP are then characterized to clarify the impact of the one-bit ADCs on the diversity order of the MIMO systems. Notably, the diversity order of the MIMO systems with one-bit ADCs is the same as the number of receive antennas only when transmitting two symbol vectors with the opposite sign. Except for this special case, the asymptotic bound of the PEP is saturated at a high signal-to-noise ratio due to the use of the one-bit ADCs. Using simulations, we verify that the analytical results coincide with the asymptotic behavior of the PEP of the MLD for MIMO systems with one-bit ADCs over Rayleigh fading channels.

5 November 2019 in IEEE Access
Verified
Journal Article
Open access
YKYun-Seong Kang
Yun-Seong Kang
MMMoonsik Min
Moonsik Min
A downlink cellular network in which base stations (BSs) are distributed according to a homogeneous Poisson point process is considered. Each BS simultaneously serves multiple single-antenna users for downlink transmission; zero-forcing beamforming (ZFBF) is used for spatial division multiplexing. Each user quantizes and feeds back the channel state information (CSI) to the transmitter to ensure that at least partial CSI is available at the transmitter. The net spectral efficiency is introduced as a main performance metric; it measures the net profit achieved by using limited-feedback-based ZFBF. This paper mainly analyzes the optimal number of feedback bits that maximizes the net spectral efficiency. The optimal number is analyzed with respect to various important system parameters by extending and generalizing previous studies. Specifically, this paper analyzes the asymptotic behaviors of the optimal number with respect to the signal-to-noise ratio (SNR) and the channel coherence time; it does so by providing close approximations to the optimal number that is generally satisfied regardless of the values of the SNR, number of users, path loss exponent, and channel coherence time. It is demonstrated by simulation that the proposed approximations are extremely close to the optimal number on the region of interest of the system parameters.