| This dissertation concerns performance improvement for OFDM communications in time-varying channel.; A signal propagating through the wireless channel usually arrives at receiver along a number of different paths, referred to as multipaths. Fast fading results from the geometry changes from transmitter to receiver in time, which is mainly due to the movement of mobile unit. Multipath propagation and fast fading result in the spreading of the signal in different dimensions. They are delay spread, Doppler spread and angular spread. These spreads have significant effects on the signal, and in most cases, they degrade the performance of conventional receivers; However, exploiting space, time and frequency dimensions simultaneously, the new receiver processing schemes enable lower BER for communications in highly time-varying channel. The proposed schemes, so-called single summation space-time-frequency processing (STFP) and double summation STFP are introduced and their properties are analyzed and compared.; As a special case of STFP, Doppler diversity schemes are discussed especially for OFDM communications. Doppler diversity for OFDM systems can be realized in both frequency and time domain. Frequency domain approach requires less complexity than time domain approach. However, it requires accurate information on the channel. Time domain approach utilizes pilot subsymbols to estimate required signal correlations. From the correlations, optimum weights for diversity combining are computed.; Underwater acoustic (UWA) communications channel is highly time-varying. Because of the low sound velocity, the inhomogeneity of medium, reflections from surface in motion, and so on, UWA channel shows high temporal and spectral variations. Therefore, the communications in underwater is quite challenging.; In this thesis, UWA channel is modeled as a time-varying filter composed of minimum-phase systems and all-pass systems. Then, a new channel simulator based on this model is designed, and it enables realistic simulations of UWA channels. The results are compared with real data recorded at sea.; An OFDM receiver processing is developed for UWA communications. In order to combat Doppler shift and spreads, it employs Doppler estimator and compensator, and time-frequency channel estimation and equalization. Its performance is demonstrated through the sea trials. |
