Software defined radios, from base-band architecture to reconfigurable antennas

Orthogonal Frequency Division Multiple Access (OFDMA) paired with Multiple Input Multiple Output (MIMO) antenna techniques provide high diversity order along with high spectral efficiency. To enhance the performance of MIMO-OFDMA systems, several algorithms are used to mitigate the link imperfections as well as to recover transmitted information. To further improve performance, directional reconfigurable antennas with beam-tilting capability can be used.;Performance gains in MIMO systems are highly dependent on the accurate representation of channel conditions. While simplified channel models are sufficient for initial performance characterizations, channel emulators provide the capability of emulating a vast variety of wireless channels in a lab environment. In this work, design and implementation of a broadband wireless channel emulator in the frequency domain is presented as an efficient and scalable alternative to the conventional Finite Impulse Response (FIR) based channel emulators for MIMO systems. It is shown that for a 3x3 MIMO array the proposed emulator results in 67% hardware reduction compared to the similarly constraint FIR-based emulator.;Furthermore, to compare and quantify the effectiveness of different algorithms, a platform supporting highly modular MIMO-OFDMA is essential. On such a platform, multiple algorithms can be used for the same purpose and their gains analyzed and compared. Also, when equipped with reconfigurable capabilities, the platform can be used as part of a Software Defined Radio (SDR) system to adapt its parameters and specs to the current state of the wireless channel. To this end, a MIMO-OFDMA platform was developed with a modular and reconfigurable architecture.;Finally, the link performance enhancement by using directional reconfigurable antennas with beam-tilting capability is analyzed. It is shown that higher directivity of directional antennas when combined with an agile direction selection scheme results in higher capacity or lower Symbol Error Rate (SER) in the system. Also, the training overhead for direction selection for such antennas is quantified. To reduce this overhead several training schemes are proposed and their capacity performances are compared with the conventional exhaustive training scheme. In particular, it is shown that using the proposed Exhaustive Training at Reduced Frequency (ETRF) consistently results in a reduction in training overhead. It is also demonstrated that further reduction in training overhead is possible using statistical or MUSIC-based training schemes.