Finite Sample Performance of Channel Estimation and Equalization for SIMO-OFDM Systems without Cyclic Prefix

• Main Authors: Mei-Chen Lee, 李美貞
• Other Authors: 余金郎
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/16209633298485851932

碩士 === 輔仁大學 === 電子工程學系 === 96 === In orthogonal frequency-division multiplexing (OFDM) systems, cyclic prefix (CP) insertion and channel estimation are indispensable to achieve coherent demodulation. However, the use of cyclic prefix causes a significant loss in bandwidth efficiency. The subspace (SS)-based blind channel estimation methods for OFDM systems without CP has been presented to gain the bandwidth efficiency. Furthermore, the block matrix scheme is also proposed to enhance the performance of SS blind channel estimation. The block SS channel estimation obtains both lower computational complexity and faster convergence speed than the SS channel estimation. In this thesis, based on the block-SS channel estimator, the adaptive equalization in the finite sample scenario is studied. We will focus on the design of adaptive equalization for OFDM systems without CP and propose various bandwidth efficient (BWE) equalizations in time domain and frequency domain, respectively. The idea of our receiver is inspired by the minimum variance (MV) approach used in adaptive beamformer. By setting unit gain on the desired signal and zero gains on the other ones, we can eliminate the intersymbol interference (ISI) and intercarrier interference (ICI). We use the Lagrange multipliers method to obtain the weight. To improve the performance in the finite sample scenario, the ISI/ICI-free matched filter and diagonal loading (DL) approach are used to reduce the perturbation of filter weight. The proposed equalization can attain a low bit error rate using a small number of samples. Its spectral efficiency and performance is better than conventional equalization.