|
|
Spatial Diversity and Time-Frequency Orthogonality for Robust Wireless Communications
Zedka, Radim ; Armada, Ana Garcia (oponent) ; Rakús,, Martin (oponent) ; Götthans, Tomáš (vedoucí práce)
This dissertation thesis contains several contributions in the field of robust wireless communications. The first part is focused on developing and analyzing a novel technique of spatial diversity based on the recently discovered concept of space-time line codes (STLCs). Unlike the already known orthogonal STLC, this design is based on a quasi-orthogonal code structure which allows the utilization of four receiver antennas instead of two antennas, which is the limit of the orthogonal STLC for a full rate transmission. The work provides a rigorous theoretical analysis of the diversity gain and asymptotic SNR gain of the new STLC and compares it to the benchmark STLCs. The second part deals with the orthogonal-time-frequency-space (OTFS), a popular multi-carrier waveform showing excellent performance in high-mobility wireless communications. The contribution lies in designing a new OTFS-like system, which - unlike the vast majority of OTFS literature - considers the power-spectral-density (PSD) one of the key design parameters. The novel waveform is named unique-word OTFS (UW-OTFS), because its outer stage is adopted from the popular UW-OFDM. UW-OTFS inherits all of its benefits, including the improved bit-error-ratio (BER) and superb PSD properties. Moreover, the main issue of UW-OFDM, i.e., the multiple-tap equalization, is no longer problematic in the OTFS context. The proposed UW-OTFS is designed to operate in linear time-variant (LTV) fractional-Doppler channels by utilizing a modification of the basis-expansion-model (BEM). Due to this method, both the channel estimation and equalization are performed by the linear minimum-mean-square-error (LMMSE) technique. The proposed system is analyzed both theoretically and numerically (using the Monte Carlo method) and it shows excellent performance in fractional-Doppler channels.
|
host ::
RSS.