Xiang Yun

Investigation on elevation beamforming for future LTE-Advanced

This paper investigates and analyzes three typical elevation beamforming scenarios which are most likely to be applied in future LTE-Advanced systems: vertical sectorization with same carrier frequency, vertical sectorization with different carrier frequency based on carrier aggregation, and user-specific elevation beamforming. Preliminary evaluation using WINNERII/WINNER+ based 3D MIMO channel modeling is carried out to compare these scenarios with the conventional system with single downtilting. It is shown that the latter two elevation beamforming scenarios can achieve good performance.

A Practical Design of Downlink Coordinated Multi-Point Transmission for LTE-Advanced

Coordinated Multi-Point (CoMP) transmission is considered for Long Term Evolution - Advanced (LTE-A) as an effective method to improve both average and cell-edge throughput. However, most of current research on CoMP with Multi-User (MU)-MIMO are under ideal assumption. In this paper, we study CoMP under realistic conditions, including limited feedback and complexity. Taking both performance and complexity into consideration, a hybrid cooperating set selection scheme is proposed. Furthermore, Channel Quality Indicator (CQI) estimation based on instantaneous precoder and CQI update based on statistical distribtion are proposed to improve CQI accuracy for Modulation and Coding Scheme (MCS) selection. Finally, the system level simulation results of CoMP with MU-MIMO are provided, which show that, compared with single cell transmission, CoMP MU-MIMO with our proposed scheme can achieve considerable performance gain in terms of both average and cell-edge throughput.

Power allocation of Dynamic Point Blanking for downlink CoMP in LTE-Advanced

Coordinated Multiple Point (CoMP) has been adopted in 3GPP LTE-Advanced to improve both the cell average and cell edge throughput. Among various downlink CoMP schemes, Dynamic Point Blanking (DPB) is simple and effective to improve the cell-edge user throughput by dynamic muting of the strongest interference of coordinated points. DPB can be used together with Coordinated Scheduling (CS) as CS/DPB and Dynamic Point Selection (DPS) as DPS/DPB to achieve more cell-edge throughput. However there exists the problem that radio resource may not be fully used due to muting. In this paper, Dynamic Point Reduced Power (DPRP) schemes are proposed to further improve the cell edge throughput. With the proposed methods, each resource block (RB) can be dynamically allocated with zero power, reduced power or a full power. Performance evaluation shows the proposed methods improve cell edge throughput by 30% compared to that of conventional single cell transmission.

An enhanced interference measurement scheme for CoMP in LTE-Advanced downlink

This paper proposes an enhanced interference measurement method that provides accurate channel quality indicators (CQIs) to improve the system performance in coordinated multi-point (CoMP) transmission and reception. In CoMP operations, user equipments (UEs) need to report different CQIs based on different interference hypotheses. Even though a variety of interference measurement schemes are proposed, they always suffer higher complexity, inaccuracy or restriction in terms of the number of available interference measurement resource (IMR) configurations. In the proposed method, phase rotations are introduced for channel state information reference signals (CSI-RSs) before mapping to the resource grid. Both theoretical analysis and simulation results show that the proposed method improves the accuracies of both signal estimation and interference measurement, and consequently yields accurate CQIs. Moreover, it can provide a large number of IMR configurations.

Investigation of Optimum Double Codebook Design for Downlink MIMO in LTE-A

Due to the limited feedback amount, the codebook and feedback design is very important. In order to achieve better tradeoff between overhead and performance, the double codebook structures are proposed and discussed, where a final precoder W is determined by two precoders W1 and W2, which are selected from two different codebooks. In this paper, we give an overview of several double codebook structures and evaluate them based on system level simulations (SLS). We consider two frameworks to determine a final precoder: W1W2 and W2W1 where W1 and W2 track long-term/wideband and short-term/subband channel properties, respectively. It is shown that under 4Tx assumption, double codebooks achieve gains compared with Rel.8 codebook when considering the multi-user (MU) MIMO and do not suffer in the case of single-user (SU) MIMO. Both W1W2 and W2W1 provide nearly the same gains over that for the Rel. 8 codebook with uniform linear array (ULA) deployment. Furthermore, W1W2 framework performs better than W2W1 framework with cross-polarized array (CPA) deployment.

Investigation on CQI Definition for CoMP in LTE-Advanced Downlink

In order to support dynamic switching among different coordinated multi-point (CoMP) schemes, the eNodeB must have channel quality indicator (CQI) information for different schemes to perform scheduling. The full feedback scheme can provide accurate CQI for different CoMP schemes but it comes at the cost of a large feedback overhead. A new CQI definition of relative interference is proposed to reduce the feedback overhead while enabling the eNodeB to obtain various types of CQI for different CoMP schemes through CQI updates. In this paper, we compare the proposed scheme with other four candidate CQI definitions, and clarify their performance impact under different traffic loads. Evaluation results show that the proposed CQI definition of relative interference works effectively under different traffic loads for different CoMP schemes.

Interference Measurement Scheme for CoMP in LTE-Advanced Downlink

Coordinated multi-point (CoMP) transmission and reception is considered to be a promising technology to improve the cell-edge user throughput by mitigating the inter-cell interference, and it is expected to be specified in long-term evolution (LTE) Release 11 in 3GPP. For CoMP, the channel state information (CSI) feedback mechanism is a key factor in achieving a better throughput performance gain. This paper proposes two interference measurement schemes for CoMP in the LTE-Advanced downlink. In the proposed schemes, the existing CSI reference signal, which is used for CSI measurement, is fully utilized considering the impact to legacy user equipment, and a sufficient number of configurations for interference measurement resources are established to facilitate operator cell-planning efforts. Based on multi-cell link level and system-level simulations, we also confirm that the proposed schemes achieve an accurate interference measurement for a single transmission point and CoMP operation.

Novel CQI update methods in CoMP transmission

This paper proposes two novel channel quality indicator (CQI) update methods that provide conservative CQIs to improve the cell edge user throughput in coordinated scheduling/beamforming (CS/CB) schemes. In CS/CB, a user usually needs to feed back channel state information (CSI) to its serving cell as the first step. However, due to the lack of knowledge regarding neighboring coordinated cells, e.g., the precoding matrix, the reported CQI is not accurate and must be updated before it can be applied. In the proposed methods, we assume that for each data stream the user receiver is tuned to match each incoming source of interference from the coordinated cell, or the precoder of the coordinated cell matches the row space of the channel from the coordinated cell to the user as much as possible. Two lower-bounds of the real CQI in actual transmission are derived and adopted. Both theoretical analysis and simulation results show the effectiveness of the proposed methods.

Multiple-CSI selection methods for CoMP downlink

Accurate channel state information (CSI) is essential for scheduling and modulation and coding scheme (MCS) selection at eNodeB. For coordinated multi-point (CoMP) transmission, multiple CSIs are necessary for eNodeB to do dynamic switch among different CoMP schemes. Due to the limitation of feedback overhead and user complexity, only N CSIs can be configured for feedback. Under such circumstance, how to select suitable multiple CSIs, and ensure accurate channel quality indicator (CQI) update become an important problem. In this paper, we propose a new two-step CSI selection method to predict proper CoMP scheme and select relevant CSI for accurate CQI update. Simulation evaluation shows that the proposed method provides better performance on both average and edge user throughput than conventional methods through appropriate CSI selection and accurate CQI update.

Multi-tier token based scheduling considering QoS and revenue in OFDMA system

Due to the propagation of smart handhold terminals, e.g. iPhone and iPad, the data traffic is growing rapidly in cellular networks. To fulfill the increased data traffic and attract more customers, the operators need to construct the network infrastructure and update their service planning, e.g. via introduction of flat-rate (FR) service plan. While along with the increase of data users, especially FR users, the operators find that the revenue and QoS (Quality of Service) will not improve accordingly. One important reason is that the over-transmitted FR services will distinctly affect the QoS of non-flat-rate (NFR) ones, and meanwhile lead very limited revenue benefit to operators. The conventional QoS scheduling researches are mainly targeting to UE (User Equipment) side issues, e.g. GBR (Guaranteed Bit Rate), instead of the revenue from operator side. To consider both UEs QoS and operational revenue, the paper firstly introduces the Token based Proportional Fairness (TPF) scheduling as the premise, and then presents the Multi-tier Token Model (MTM) for FR and NFR differentiation. After that, the two-stage PF factor updating method for revenue optimization is proposed. The system evaluations show that the proposed solution of revenue oriented scheduling can properly distribute the radio resource for higher income tendentiousness, and in the meantime fulfill the GBR requests for both FR and NFR services.