Novel SISO Based Blind Channel Estimator And Long-Range Predictor For Wireless OFDM Systems

Novel SISO based blind channel estimator is derived in this paper for OFDM systems. The estimator consists of two parts: channel magnitude frequency response estimator, and channel phase frequency response estimator. After exploiting the second order instantaneous moment of the received signals in frequency domain, the proposed complex exponential expansion model is introduced to identify the magnitude frequency response of OFDM channel with few parameters. Then, the phase estimator can further determine the phase of frequency response of OFDM channel. Through the whole algorithm, only a small amount of computation is involved. Therefore, the algorithms can track the channel continuously and are amenable to real applications. The algorithm is tested with simulations and also compared with other methods.Wireless channel predictor will be the key section of the next generation wireless network. In this thesis a frequency domain channel predictor for OFDM system is discussed. Reliable prediction is feasible about 15ms ahead for the Doppler frequency offset of 100Hz. Compared with time domain predictor, the computational complexity of the proposed scheme is largely reduced, because it replaces convolution calculation with transposition multiplication. While the OFDM system is involved, our predictor can also be low hardware complexity. In handset side, digital signal preceding resource islimited. In many cases, the FFT sections are solidified for faster or saving cost. If the fast furrier transform (FFT) for OFDM modulator (or demodulator) can be reused in our predictor, the structure we have developed provides a priority scheme for the next generation mobile (wireless) OFDM network adaptive transmission. It is observed, from the simulation results, that the predictor with one step (block) offset bears better performance.