Abstract :
[en] The utilization of multi-antenna transmitters relying on full frequency reuse has proven to be an effective strategy towards fulfilling the constantly increasing throughput requirements of wireless communication systems. As a consequence, in the last two decades precoding has been a prolific research area, due to its ability to handle the interference arising among simultaneous transmissions addressed to different co-channel users. The conventional precoding strategies aim at mitigating the multi-user interference (MUI) by exploiting the knowledge of the channel state information (CSI). More recently, novel approaches have been proposed where the aim is not to eliminate the interference, but rather to control it so as to achieve a constructive interference effect at each receiver. In these schemes, referred to as symbol-level precoding (SLP), the data information (data symbols) is used together with the CSI in the precoding design, which can be addressed following several optimization strategies.
In the context of SLP, the work carried out in this thesis is mainly focused on developing more advanced optimization strategies suitable to non-linear systems, where the per-antenna high-power amplifiers introduce an amplitude and phase distortion on the transmitted signals. More specifically, the main objective is to exploit the potential of SLP not only to achieve the constructive interference at the receivers, but also to control the per-antenna instantaneous transmit power, improving the power dynamics of the transmitted waveforms. In fact, a reduction of the power variation of the signals, both in the spatial dimension (across the different antennas) and in the temporal dimension, is particularly important for mitigating the non-linear effects.
After a detailed review of the state of the art of SLP, the first part of the thesis is focused on improving the power dynamics of the transmitted signals in the spatial dimension, by reducing the instantaneous power imbalances across the different antennas. First, a SLP per-antenna power minimization scheme is presented, followed by a related max-min fair formulation with per-antenna power constraints. These approaches allow to reduce the power peaks of the signals across the antennas. Next, more advanced SLP schemes are formulated and solved, with the objective of further improving the spatial dynamics of the signals. Specifically, a first approach performs a peak power minimization under a lower bound constraint on the per-antenna transmit power, while a second strategy minimizes the spatial peak-to-average power ratio.
The second part of this thesis is devoted to developing a novel SLP method, referred to as spatio-temporal SLP, where the temporal variation of the transmit power is also considered in the SLP optimization. This new model allows to minimize the peak-to-average power ratio of the transmitted waveforms both in the spatial and in the temporal dimensions, thus further improving the robustness of the signals to non-linear effects. Then, this thesis takes one step further, by exploiting the developed spatio-temporal SLP model in a different context. In particular, a spatio-temporal SLP scheme is proposed which enables faster-than-Nyquist (FTN) signaling over multi-user systems, by constructively handling at the transmitter side not only the MUI but also the inter-symbol interference (ISI). This strategy allows to benefit from the increased throughput provided by FTN signaling without imposing additional complexity at the user terminals. Extensive numerical results are presented throughout the thesis, in order to assess the performance of the proposed schemes with respect to the state of the art in SLP. The thesis concludes summarizing the main research findings and identifying the open problems, which will constitute the basis for the future work.
