Channel Estimation and Equalization for OFDM Wireless System with Medium Doppler Spread

A new channel estimator ML+SOPI is proposed for medium time-variant wireless channel, where ML denotes maximum likelihood used to estimate the CIR of pilot sequence (PS) whereas second-order polynomial (SOP) interpolation (SOPI) is to compute channel matrix of each symbol by these estimated CIRs. Then, three schemes are formed as follows: scheme A , an union of ML+SOPI and one-tap equalizer (OTE), where interpolation is based on symbol level and CIR viewed as constant; scheme B, a combination of ML+SOPI and zero- forcing equalizer (ZFE), where interpolation is conducted on sampling-point level and CIR approximated as linear; scheme C, an union of ML+SOPI and Jeon equalizer with the same assumptions as scheme B. The complexity of ZFE in scheme B is reduced by Neumann expansion and IFFT/FFT, which has far lower complexity than Jeon equalizer for long cyclic prefix(>10). The simulated results in mobile channel are as follows: when normalized Doppler spread (NDS) approaches 0.1, schemes B and C realize approximately the same BER performance, and perform far better than scheme A.Thus, scheme B is preferred for practical applications.