Improved DFT-based channel estimation for OFDM systems over multipath channels

Xiaoqing Wang,Jintao Wang,Zhaocheng Wang,Zhixing Yang,Jian Song
DOI: https://doi.org/10.1145/1815396.1815532
2010-01-01
Abstract:To efficiently suppress the noise through a time-domain threshold value after the least square (LS) estimation, the discrete Fourier transform (DFT) based channel estimation (CE) is usually carried out in practical orthogonal frequency division multiplexing (OFDM) systems for its inherent noise immunity. In this paper, an improved DFT based CE method for OFDM systems is proposed. Taking advantage of the wavelet decomposition, the proposed method obtains an optimal threshold value according to the minimum mean-square error (MMSE) optimality criterion. Both theoretical analysis and computer simulation show that, the proposed method not only reduces the mean-square error (MSE) without MSE floor compared with the conventional methods, but also improves the symbol error rate (SER) performance by being more close to that of the ideal known-channel case over both static and dynamic multipath channels.
What problem does this paper attempt to address?