Kengo Ando

• This dissertation contributes to three distinct focuses of BF design as: BF for flexible connectivity, BF for enhanced connectivity, and BF for over-the-air-computating (AirComp). The first objective is the “BF for flexible connectivity”. In order to realize flexibility for preserving connectivity regardless of the user position in the coverage area, the recent cell-free MIMO (CF-MIMO) system is considered. For the BF design, a flexible design is proposed, which is directly adaptable not only for both uplink (UL) and downlink (DL) communication modes but also for both under-loaded and over-loaded scenarios. As for the second focus, termed “BF for enhanced connectivity,” a novel BF design compatible with three distinct power allocation schemes is proposed. For the sake of connectivity enhancement, a DL MIMO-rate splitting multiple access (RSMA) system is considered. Even though the proposed BF design is assumed to be used only for under-loaded or fully-loaded scenarios, the computational complexity required to design BF is significantly less than that of the state-of-the-art (SotA) alternative. Finally, the ”BF for Over the Air Computing” is considered for envisioning the realization of integrated AI and communication. In this focus, novel receiver (RX) BF designs compatible with uniform-forcing (UF) precoding for a multi-user UL multiple-input single-output (MISO)-AirComp system are considered for higher performance or lower complexity. Toward higher performance design, while the proposed design sacrifices computational complexity in the BF, the resulting AirComp has a lower mean square error(MSE) performance. On the other hand, the proposed BF design for lower complexity realizes equivalent MSE performance achieved by the high-performance BF design with significantly lower complexity thanks to the combination of recent convex optimization and Bayesian optimization (BO) methods.