Wireless Ofdm Systems, Channel Estimation And Equalization Algorithm Research,

Future wirelesss communication shows wider bandwidth, higher carrier frequency andhigher mobile speed. The blocking transmission technique such as OFDM (OrthogonalFrequency Division Multiplexing) can efficiently cancel multipath interference caused byincreasing signal bandwidth, achieve high bandwidth efficiency and offerfrequency-domain equalizer with low complexity. Higher carrier frequency and fastermobile speed leads to ICI (Inter Channel Interference). When the NDS (NormalizedDoppler Spread) is larger than 0.01, the constant channel model will not hold, so a linear ornonlinear model for the channel is required to suppress ICI. In addition, quite a fewwireless channels are sparse. If the sparsity can be further exploited, the effect of channelnoise can be efficiently reduced. This implies that the performance of the system can besubstantially improved.The paper consists of three parts: sparse channel estimation, time-variant channelestimation and time-various channel equalization. The main contributions of the paper areas follows:a) We propose a joint channel estimator based on ML (Maximum Likelihood) rule,which can estimate CIR and noise variance simultaneously. Based on this, combining twoestimated paraneters can yield a threshold-based ML channel estimator, called as IML(Improved ML). By the simulation in the HF/MF channel, we obtain: IML shows 2dBSNR gain in BER performance over the the traditional ML estimator, its performance isvery close to that of the ideal estimator. Compared with other estimators, the IML has thelower complexity and more simple structure.b) We propose an estimator based on ML rule for 0.01＜NDS＜0.1, which usestime-domain pilot sequence. This estimator is implemented as follows: first, we can get theCIR estimation corresponding to the pilot sequence, then, we interpolate the CIR of datasymbols by second-order polynomial interpolation (SOPI) using the estimated CIR in thefirst step. By the simulation in the mobile urban channel, it follows that the performance ofthe SOPI is suprior to linear interpolation.c) We propose an implified ZF (Zero-Forcing) equalizer with lower complexity for themedium time-variant channel. This equalizer combines the Nuemann series and theFFT/IFFT technique, so it can make ZF equalizer very simple. By the simulation in themobile urban channel, the performance of the equalizer proposed by us approaches that oforiginal ZF equalizer.