LTE-A 단말기에서의 SIC 기반 채널 추정을 통한 셀간 간섭 제거 기법

Abstract

국제 통신 연합-무선 통신 섹터(International Telecommunication Union-Radio communication Sector: ITU-R)는 4세대 무선 통신 시스템으로 Long Term Evolution-Advanced(LTE-A)을 채택하였다. LTE-A는 다중 송신 안테나와 수신 안테나를 사용하는 다중 입력 다중 출력(Multiple-Input Multiple-Output: MIMO)과 다운링크에서 리소스 할당 효율을 높이기 위해 직교 주파수 분할 다중 접속(Orthogonal Frequency-Division Multiple Access: OFDMA)을 사용한다. 셀룰러 통신 인프라 구조의 셀 경계 지역에서 인접 셀로부터 주파수 채널 간섭에 따른 커다란 인접 셀 간섭(Inter-Cell Interference: ICI)과 신호 감쇄와 인접 셀 간섭에 따른 불완전한 채널 상태 정보(Channel State Information: CSI)가 사용자 단말기(User Equipments: UEs)에 주로 나타나는데, 수신기는 이러한 불완전한 채널 상태 정보와 셀 경계에서 이웃 셀들로부터 오는 강한 간섭에 의해 신호 대 간섭 잡음 비(Signal-to-Interference-and-Noise Ratio: SINR)가 낮아지기 때문에 수신기의 성능이 심각하게 나빠진다. 셀 경계에서 낮은 정보 처리량은 Long Term Evolution(LTE) 시스템의 심각한 문제 중 하나이다. 그러므로 본 논문에서는 수신 성능을 향상시키기 위하여 2가지 방안을 제안하였고, 제안한 기법들을 검증하기 위해 링크 시뮬레이터를 구현하였다. 이 논문의 세 가지 특징을 다음과 같이 설명한다. 첫째, 최소 평균제곱오차(Minimum Mean Square Error: MMSE) 여과기를 사용하는 시스템에서 MMSE 성능을 개선하기 위해 간섭 셀의 채널 추정 정보를 사용하여 정확한 공분산 추정 방법을 제안한다. LTE 표준화를 담당하는 Third Generation Partnership Project(3GPP)의 수신 성능 개선을 위해 활동하는 Working Group 4(WG4)에서 Work Item(WI)으로 진행했던 진화된 수신기는 낮은 SINR 상황에서 정보 처리량을 높이기 위해 인접 셀 간섭 제거(Inter-Cell Interference Cancellation: ICIC) 기법으로 간단한 구조의 최소 평균제곱오차(Minimum Mean Square Error: MMSE) 여과기를 사용한다. MMSE 여과기에서는 성능을 좌우하는 인자 중 하나인 간섭을 추정하여 제거함으로써 SINR을 높인다. 이때 추정된 간섭이 실제 간섭과 유사할수록 성능이 향상되기 때문에 이 논문에서는 간섭과 유사한 인자 추정을 위해 간섭 셀의 채널 추정 정보를 사용하여 정확한 공분산 추정 방법을 제안하고 그 결과 기존 ICIC 기법보다 성능이 향상된다. 둘째, 채널 추정을 정확하게 하기 위해 채널 추정 기법에 간섭을 많이 제거할 수 있는 SIC 기법을 추가한다. ICIC 기법의 성능은 채널 추정의 정확도에 많은 영향을 받으므로 더 정확한 채널 추정이 필요하다. 그러나 셀 가장자리에서 인접 셀 신호의 크기가 상당히 클 때 UE에게 인접 셀 신호는 간섭으로 작용하므로 기존 기준 신호(Reference Signals: RS) 기반의 채널 추정은 정확도가 상당히 떨어진다. 순차적 간섭 제거(Successive Interference Cancellation: SIC) 기법은 우선 간섭 신호를 복조한 후 수신신호에서 복조된 신호를 제거함으로써 간섭을 많이 제거할 수 있는 방법이므로 정확한 채널 추정을 위해서 새로운 채널 추정 기법에 SIC 기법을 사용한다. SIC 기반의 채널 추정을 사용하기 위해 서빙 셀의 송신 타이밍과 인접 셀의 송신 타이밍이 동일하다고 가정한다. 실제로 많은 기지국들이 이미 이웃 셀과 전송 동기를 맞춰서 사용하고 있다. 동일한 기지국으로부터 신호가 송신되기 때문에 송신 타이밍은 쉽게 동기화될 수 있다. 그러므로 이 논문에서는 Long Term Evolution-Advanced(LTE-A) UE가 셀 경계에 위치할 때 채널 추정을 정확하게 하기 위해 SIC를 적용한 후 RS로 채널 추정하는 방법을 제안하고, 정확한 채널 추정 값과 정확한 채널 추정으로 도출된 MMSE 여과기의 인자를 통해 수신기의 성능을 약 5.5 dB 이상 향상시킨다. 셋째, 이 논문에서 제안한 SIC 기반 채널 추정을 통한 셀간 간섭 제거 기법의 성능을 분석하기 위해 LTE-A 기반의 링크 레벨 시뮬레이터(Link Level Simulator: LLS)를 설계하고 구현한다. LLS는 3GPP에서 정의하고 있는 여러 종류의 MIMO 전송 기술, 패킷 스케줄링 등을 적용하고, 두 셀에서 2계층 구조를 가정하여 실제적인 환경을 구성할 수 있도록 구현한다. 이를 이용하여 LTE-A 시스템에서 일반적인 채널 추정과 MMSE 여과기를 구현하고, 제안한 SIC 기반의 채널 추정과 MMSE 여과기 인자 추정 방법을 적용한다. 제안한 SIC 기반 채널 추정을 사용한 인접 셀 간섭 제거 기법의 복잡도가 얼마나 증가하는지 분석하기 위해서 수신 방식에 따른 곱셈 연산 수를 계산하고, 구현된 LLS 모델을 통해, LTE-A 시스템에서 제안된 기법의 MSE와 사용자 처리량에 대한 성능을 분석한다. 그 결과 SIC 기반 채널 추정을 사용한 제안한 ICIC 수신기는 기존의 수신기보다 곱셈 연산 수가 약 7.2배 증가한다. 그러나 EPA 채널에서 제안한 SIC 기반 채널 추정의 MSE는 기존 채널 추정의 MSE보다 약 1/9로 줄어든다. 그리고 EPA 채널에서 제안한 SIC 기반 채널 추정을 사용한 ICIC 수신기는 기존 수신기보다 6.0 dB 성능 이득을 얻는다. 그러므로 제안한 SIC 기반 채널 추정을 사용한 ICIC 수신기는 LTE-A 시스템의 성능 향상을 위해 사용될 수 있다.|Long Term Evolution-Advanced (LTE-A) was adopted by International Telecommunication Union-Radio communication Sector (ITU-R) as the fourth-generation wireless communications system in January 2012. LTE-A supports Multiple-Input Multiple-Output (MIMO), which uses multiple transmit and receive antennas to improve frequency efficiency, and Orthogonal Frequency-Division Multiple Access (OFDMA) to provide high flexibility in resource allocation for the downlink. At cell boundaries of cellular communication infrastructure, large Inter-Cell Interference (ICI) due to frequency channel interference from neighboring cell and Imperfect Channel State Information (CSI) due to signal power degradation and ICI appear mainly to User Equipments (UEs). At cell boundaries, Signal-to-Interference-plus-Noise Ratio (SINR) is low because of the strong interferences from neighboring cells and the receiver performance is severely degraded. Low data throughput at cell boundaries is one of the serious bottlenecks of an Long Term Evolution (LTE) system. Therefore, two methods are proposed to improve the receive performance and a Link Level Simulator (LLS) is implemented to confirm the proposed methods in this thesis. The three features of this thesis are described as follows. First, the exact covariance estimation scheme is proposed using estimated channel response of interference cell to improve performance of MMSE filter. In low SINR, advanced receiver that proceeds to Work Item (WI) in Working Group 4 (WG4) that activities to improve performance of Third Generation Partnership Project (3GPP) that progresses LTE standard Working Group 4 (WG4) uses simple Minimum Mean Square Error (MMSE) filter as the Inter-Cell Interference Cancellation (ICIC) scheme to improve the throughput. The estimated interference is applied to the parameter of the MMSE filter. MMSE filter estimates and reduces interference that is one of the parameters which influence the performance. Then, SINR is high. If the parameter of MMSE filter is more similar to interference then performance is improved. Therefore, in this dissertation, the accurate parameter estimation method is proposed to improve performance. Second, SIC scheme that reduces much interference is applied to accurate channel estimation. To improve the performance of the receiver, more accurate channel estimation is required since the performance of ICIC is highly dependent on the qualities of channel and parameter estimations. However, at cell boundaries, the accuracy of the channel estimation based on Reference Signals (RSs) is reduced by strong interference when signal power of neighboring cell is very strong. SIC scheme is very effective to reduce the strong signal of the neighboring cell because demodulated signal is reduced from receive signal after the interference signal is demodulated. Therefore SIC scheme is used for proposed channel estimation scheme. With the SIC-based channel estimation, it is assumed that the transmission timings of neighboring cells are aligned. In real deployments, many base stations are already deployed after aligning the transmission timings of neighboring cells. Furthermore, in a sectored cell, the transmission timings can be easily aligned since signals are transmitted from the same base station. Therefore, in this dissertation, channel estimation by RS using Successive Interference Cancellation (SIC) is presented to enhance the quality of channel estimation at cell boundaries. The receiver performance is improved about 5.5 dB or more through accurate estimated channel response and the parameter of MMSE filter estimated with the accurate estimated channel response. Third, a Link Level Simulator (LLS) based on LTE-A system is designed and implemented for performance analysis of proposed ICIC scheme with SIC-based channel estimation. The LLS applies a packet scheduler, several MIMO modes and feedback types defined by the Third Generation Partnership Project (3GPP), and is implemented in the real environment that is assumed as two-tier and two-cell configuration. By using this, the conventional channel estimation and MMSE filter are implemented in LTE-A system. And the proposed SIC-based channel estimation and a method that estimates the parameter of MMSE filter are applied. The number of multiplication according to receive system is calculated to analyze the computational complexity of the proposed ICIC scheme with SIC-based channel estimation. Through the LLS model, the performances about Mean Square Error (MSE) and user throughput of the proposed scheme are analyzed. As a result, the number of multiplications by the proposed ICIC scheme with SIC-based channel estimation increases about 2.6 times more than that of multiplication of the conventional receiver. However, in Extended Pedestrian A model (EPA) channel, it can observe that the MSE of the presented SIC-based channel estimation is about 1/9 of that of the conventional channel estimation. In EPA channel, compared with the conventional receiver, it is possible to achieve about 6.0 dB performance gain with the ICIC receiver with SIC-based channel estimation. Therefore, the proposed ICIC scheme with SIC-based channel estimation can be used to improve the performance of LTE-A systems.; Long Term Evolution-Advanced (LTE-A) was adopted by International Telecommunication Union-Radio communication Sector (ITU-R) as the fourth-generation wireless communications system in January 2012. LTE-A supports Multiple-Input Multiple-Output (MIMO), which uses multiple transmit and receive antennas to improve frequency efficiency, and Orthogonal Frequency-Division Multiple Access (OFDMA) to provide high flexibility in resource allocation for the downlink. At cell boundaries of cellular communication infrastructure, large Inter-Cell Interference (ICI) due to frequency channel interference from neighboring cell and Imperfect Channel State Information (CSI) due to signal power degradation and ICI appear mainly to User Equipments (UEs). At cell boundaries, Signal-to-Interference-plus-Noise Ratio (SINR) is low because of the strong interferences from neighboring cells and the receiver performance is severely degraded. Low data throughput at cell boundaries is one of the serious bottlenecks of an Long Term Evolution (LTE) system. Therefore, two methods are proposed to improve the receive performance and a Link Level Simulator (LLS) is implemented to confirm the proposed methods in this thesis. The three features of this thesis are described as follows. First, the exact covariance estimation scheme is proposed using estimated channel response of interference cell to improve performance of MMSE filter. In low SINR, advanced receiver that proceeds to Work Item (WI) in Working Group 4 (WG4) that activities to improve performance of Third Generation Partnership Project (3GPP) that progresses LTE standard Working Group 4 (WG4) uses simple Minimum Mean Square Error (MMSE) filter as the Inter-Cell Interference Cancellation (ICIC) scheme to improve the throughput. The estimated interference is applied to the parameter of the MMSE filter. MMSE filter estimates and reduces interference that is one of the parameters which influence the performance. Then, SINR is high. If the parameter of MMSE filter is more similar to interference then performance is improved. Therefore, in this dissertation, the accurate parameter estimation method is proposed to improve performance. Second, SIC scheme that reduces much interference is applied to accurate channel estimation. To improve the performance of the receiver, more accurate channel estimation is required since the performance of ICIC is highly dependent on the qualities of channel and parameter estimations. However, at cell boundaries, the accuracy of the channel estimation based on Reference Signals (RSs) is reduced by strong interference when signal power of neighboring cell is very strong. SIC scheme is very effective to reduce the strong signal of the neighboring cell because demodulated signal is reduced from receive signal after the interference signal is demodulated. Therefore SIC scheme is used for proposed channel estimation scheme. With the SIC-based channel estimation, it is assumed that the transmission timings of neighboring cells are aligned. In real deployments, many base stations are already deployed after aligning the transmission timings of neighboring cells. Furthermore, in a sectored cell, the transmission timings can be easily aligned since signals are transmitted from the same base station. Therefore, in this dissertation, channel estimation by RS using Successive Interference Cancellation (SIC) is presented to enhance the quality of channel estimation at cell boundaries. The receiver performance is improved about 5.5 dB or more through accurate estimated channel response and the parameter of MMSE filter estimated with the accurate estimated channel response. Third, a Link Level Simulator (LLS) based on LTE-A system is designed and implemented for performance analysis of proposed ICIC scheme with SIC-based channel estimation. The LLS applies a packet scheduler, several MIMO modes and feedback types defined by the Third Generation Partnership Project (3GPP), and is implemented in the real environment that is assumed as two-tier and two-cell configuration. By using this, the conventional channel estimation and MMSE filter are implemented in LTE-A system. And the proposed SIC-based channel estimation and a method that estimates the parameter of MMSE filter are applied. The number of multiplication according to receive system is calculated to analyze the computational complexity of the proposed ICIC scheme with SIC-based channel estimation. Through the LLS model, the performances about Mean Square Error (MSE) and user throughput of the proposed scheme are analyzed. As a result, the number of multiplications by the proposed ICIC scheme with SIC-based channel estimation increases about 2.6 times more than that of multiplication of the conventional receiver. However, in Extended Pedestrian A model (EPA) channel, it can observe that the MSE of the presented SIC-based channel estimation is about 1/9 of that of the conventional channel estimation. In EPA channel, compared with the conventional receiver, it is possible to achieve about 6.0 dB performance gain with the ICIC receiver with SIC-based channel estimation. Therefore, the proposed ICIC scheme with SIC-based channel estimation can be used to improve the performance of LTE-A systems.