OFDM系统上行的基于干扰删除的改进信道估计

本文提出了一种正交频分复用(OFDM)系统中基于码分复用(CDM)导频的改进的信道估计方法以降低导频间的非正交性干扰。在上行链路中,若不同用户的导频信号是非正交的,信道估计会受到干扰影响,从而系统性能降低。本文提出将干扰删除引入信道估计,以提高系统性能。

doc.xuehai.net

AN IMPROVED CHANNEL ESTIMATION METHOD WITH INTERFERENCE

CANCELLATION FOR OFDM SYSTEM IN UPLINK

HUANG Chen, ZHANG Jianhua, SUN Feifei

Wireless Technology Innovation Institute (WTI), Telecommunications Engineering Department,

Beijing University of Posts and Telecommunications #92 86-010-62283600(307), huangchensky@gmail.com

ABSTRACT

In this paper, an improved channel estimation method based on code division multiplexing (CDM) pilot is pro-posed to mitigate interference among pilot sequences for orthogonal frequency division multiplexing (OFDM) sys-tems in uplink. If pilot sequences of different users are not orthogonal, channel estimation suffers from inter-sequence interference in uplink, thus system performance degrades significantly. Interference cancellation is intro-duced into channel estimation to solve this problem in this paper. Simulation results show that our proposed methods have much lower complexity and comparable perform-ance with conventional decorrelating detector.

Key word: channel estimation, CDM, OFDM, interference cancellation

1. INTRODUCTION

Orthogonal frequency division multiplexing (OFDM) is regarded as a promising solution for high-speed transmis-sion of next-generation wireless communication systems in multipath fading environments [1]. WiMax and 3GPP LTE both adopt OFDM-based technologies, such as Orthogonal Frequency Division Multiple Access (OFDMA), Single Carrier-Frequency Division Multiple Access (SC-FDMA) as their multiple access technologies [2][3].

Channel estimation is a key technology of OFDM sys-tems. Bit error rate (BER) performance deteriorates greatly with the increase of the error of channel estimation, thus the system performance degrades severely. In the uplink transmission, in particular, it needs to estimate dif-ferent channels for different users, which brings new chal-lenges to channel estimation. In order to distinguish dif-ferent users, orthogonality of pilots for different users is required. The orthogonality can be achieved in frequency domain (Frequency Division Multiplexing, FDM pilot), in time domain (Time Division Multiplexing, TDM pilot), or in code domain (CDM pilot) [4][5][6].

The sequences with good auto-correlation and cross-correlation property are used for CDM pilot, such as CA-ZAC sequence, GCL sequence, and so on. According to [3], each cell is allocated with different CAZAC se-

quences with low cross-correlation; and different cyclic shifts of a base CAZAC sequence are assigned to different users for multiple access in the same cell. The amount of cyclic delay Δ should be larger than the maximum channel delay (usually equal to length of CP) plus practical user’s timing error margin. Otherwise, the Node B can not recog-nize users’ signal correctly. However the users near the cell border will face a “dominant interferer” from the users of neighbor cells, especially the ones near the border of neighbor cell. Besides, if necessary, more than one base CA-ZAC sequence can be applied in each cell to support enough simultaneous users and this will also lead to inter-sequence interference, thus degrade the performance of channel esti-mation too. To eliminate this interference, decorrelating de-tector can be used, but it requires inverse matrix calculation and high complexity. When the number of users is larger thanNΔ, the equation group of decorrelating is underde-termined, and then decorrelating detector can not be used. Interference cancellation (IC) has been widely used in sig-nal detection for many years [7][8][9][10]. This paper intro-duces parallel interference cancellation to channel estimation in order to reduce the inter-sequence interference. One prominent difference between IC in signal detection and IC in channel estimation is that unlike modulation signals, real channel response does not have fix amplitude and phase set, and detect-based non-linear interference reconstruct can’t apply to channel estimation. The main idea of our proposed methods is mitigating pilot pollution by using the accurate channel information of last frame and parallel (PIC) interfer-ence cancellation. Our proposed methods do not need inverse matrix computation, and has lower computational complex-ity. Besides unlike decorrelating detector, there is little limi-tation on the number of unknown channels, namely, when the number of unknown channel is larger than NΔ, our method still works, of course with a little deterioration in performance.

The rest of this paper is organized as follows. System model and traditional channel estimation is described in sec-tion Ⅱ. Section Ⅲ is devoted to our proposed methods. Simulation parameters and results are given in section Ⅳ, which show MSE (mean square error) improvement of the proposed methods compared to conventional method. Sec-

第1页下一页