Gibum Kim

Performance of linear precoding and user selection in IEEE 802.11ac downlink MU-MIMO system

IEEE 802.11ac is a next generation wireless local area network (WLAN) standard employing multi-user multiple-input multiple-output (MU-MIMO) scheme for the first time. Since precoding and user selection schemes influence the performance of MU-MIMO transmission, accurate channel state information (CSI) is required to select users and obtain precoding matrix. In this paper, we evaluate the performance of linear precoding schemes such as channel inversion (CI) and block diagonalization (BD) with limited CSI by using compressed feedback. We also describe the greedy user selection metrics and linear precoding schemes of low complexity. Our results compare the achievable sum-rate between the user selection metrics and provide the packet error rate (PER) performance under the realistic channel models.

Adaptive frame size estimation using extended Kalman filter for high-stressed WLANs

Demands for high throughput and stable service quality are increasing. Frame aggregation mechanisms in IEEE 802.11n wireless local area networks (WLANs) can provide improved throughput, but the effect of A-MSDU decreases significantly in error-prone channels. Therefore, adaptive frame size estimation (FSE) depending on the channel condition is required to maintain the improved throughput. In this paper, we proposed frame error rate (FER) based FSE scheme in error-prone and time-varying channel such as a high-stressed network. A tight FER bound is derived to obtain instantaneous link condition, and extended Kalman filter (EKF) is used to estimate frame size optimally for next transmission with current channel information. Our simulation results show that the proposed FSE scheme improves the throughput two times higher than a nonadaptation approach in high-stressed network condition.

Quantization error reduction scheme for hybrid beamforming

In a multiple-input multiple-output (MIMO) system of a large number of antennas, the requirement of an analog to digital converter (ADC) in each antenna string causes high implementation cost and excessive signal processing. Hence, a hybrid beamforming has been proposed in the literature as a solution. However, still quantization errors introduced at the analog beamformer part can lead to performance degradation. This paper is to present a quantization error reduction scheme for a hybrid beamformer in a 2 by 2 MIMO system. The key idea is to sacrifice the data rate from a full rate to a half rate, estimate all user data in a digital baseband, and then restore the received signal at each antenna string. By using the restored received signal and the residual (which is calculated in the estimator), quantization error can be reduced. Simulation results verify that the proposed scheme can improve the bit error rate performance with a resolution of smaller number of bits.

Asymptotic Sum Rate in Spatially Correlated Two-Cell Channel With User Scheduling

Coordinated beamforming (CBF) provides improved sum rate with less overhead and computational complexity, and its theoretical performance analysis is mostly finished in a spatially uncorrelated environment. In this paper, we investigate the effect of spatial correlation on the sum rate of a two-cell multiuser downlink channel. We analyze the scaling laws of the sum rate with a large number of users. To improve the performance, different roles are assigned to the two cells: One is a master cell, and the other is a slave cell. The base station (BS) of the master cell requests the interference cancellation to the slave cell, and the BS of the slave cell performs precoding to cancel the interference influencing over the master cell, although the achievable rate of the slave cell is decreased. The decision of the master and slave cells can be affected by the number of users and the degree of spatial correlations. From computer-based simulations, we show that our proposed algorithm can achieve a higher sum rate than other previous approaches with lower overhead.

A low complexity user selection scheme in downlink MU-MIMO system

In this paper, we describe a low complexity greedy user selection scheme for multiuser MIMO systems. We propose a new metric which has significantly reduced computational complexity and improved performance compared to Frobenius norm. The approximation of projection matrix is applied to reduce the number of singular value decomposition (SVD) operations. We analyze the computational complexity of metrics, and verify the reduced complexity by numerical evaluation. Our simulation result shows that the proposed metric can achieve the improved sum rate as the number of user antenna increases.

Improved link adaptation using truncated channel inversion for MU-MIMO systems

In this work, we propose a link adaptation scheme called truncated channel inversion (TCI) as an effective power allocation strategy for block adaptive modulation, coding, and spatial mode (AMCS). Since water-filling based power allocation is no longer optimal to minimize bit error rate (BER) in block modulation system, block AMCS is considered as a practical link adaptation scheme. We first investigate the BER bound for multiuser multiple-input multiple-output bit-interleaved coded orthogonal frequency division multiplexing (MU-MIMO BIC-OFDM), and the optimal power allocation solution that minimizes the bound is derived. By showing that the solution from the BER minimization and the result from TCI are equal, we verify the suitability of TCI for MU-MIMO BIC-OFDM system with block modulation. Then, we propose a block AMCS algorithm that integrates TCI, and selects the optimal MCS and the corresponding power allocation which maximize the throughput under a packet error rate (PER) constraint. Improved transmission performance of the proposed algorithm is verified through simulation results.

Machine-Learning-Based Link Adaptation for Energy-Efficient MIMO-OFDM Systems

Recent wireless communication trends have emphasized the importance of energy-efficient transmission. In this paper, link adaptation with machine learning mechanism for maximum energy efficiency in multiple-input multiple-output orthogonal frequency division multiplexing(MIMO-OFDM) wireless system is considered. For reflecting frequency-selective MIMO-OFDM channels, two-dimensional capacity(2D-CAP) feature space is proposed. In addition, machine-learning-based bit and power adaptation(ML-BPA) algorithm that performs classification-based link adaptation is presented. Simulation results show that 2D-CAP feature space can represent channel conditions accurately and bring noticeable improvement in link adaptation performance. Compared with other feature spaces, including ordered postprocessing signal-to-noise ratio(ordSNR) feature space, 2D-CAP has distinguished advantages in either efficiency performance or computational complexity.