Research On Channel Estimation For OFDM System With Wavelet De-noising

OFDM can eliminate the ISI caused by the time dispersion of wireless channel.Due to its high-rate transmission capability, high bandwidth efficiency, robustnessto resist multi-path fading and delay and its approachability to other multiple accesstechnologies, OFDM has been widely used in xDSL system, WLAN IEEE802.11a/g/n (WiFi) and wireless broadband access standard IEEE802.16(Wi-Max).The MIMO-OFDM technology which is the combination of OFDM andMIMO will be applied in B3G (Beyond3G) and4G.For better performance, the receiver of OFDM system usually utilizes coherentdemodulation. As the coherent detection and space-time decoder both need accuratechannel information, the accuracy of the channel estimation affect the performanceof the entire communication system and channel estimation becomes an importantpart of the receiver in OFDM system. Pilot-tone aided channel estimation is able totrack the change of the channel, and can strike a better balance between real-timeand computational complexity.The major features of wireless fading channel and the basic principles ofOFDM and MIMO technologies which include the model, the implementation, thekey technologies and the relative merits are introduced in the first section.Pilot-aided channel estimation is applied widely in practical communicationsystem due to its real-time and low computational complexity. This paper analyzesthe application scenes of different pilot patterns. This section discusses thedifference between two classical channel estimation methods, LS method andMMSE method, and compares the performance of different interpolation methods.Reducing the influence of noise on the channel estimation can improve theestimation precision of the LS algorithm, so the wavelet de-noising is introducedinto channel estimation of OFDM system. This paper proposes a method withwavelet de-noising for OFDM channel estimation. After getting the CFR on thepilot tone, the proposed method process the CFR with wavelet de-noising, and atlast the CFR of the whole band is obtained via interpolation.Finally, the simulations on the scenes of SISO-OFDM and MIMO-OFDMdemonstrate that the wavelet de-noising channel estimation method proposed by thispaper can reduce the effect of noise and ICI in the LS channel estimation andexhibits an improved performance.