SOLID: System-wide Operation via Learning In-device Dissimilarities

The key challenge in wireless communications is increasing the data rate at the device. To achieve that, modern wireless communication standards employ so-called MIMO (multiple-input multiple-output) techniques that allow parallelizing the data transmission over multiple streams using multiple device antennas. However, the growing diversity of the device types (not only handsets but also aerial vehicles, automobiles, robots, etc.) and high mobility challenge the current MIMO design for 5G and beyond networks. This project is a cooperation among wireless communications experts from North Carolina State University (NC State) and Tampere University (TAU). It develops machine learning (ML)-based solutions to empower devices to learn optimal antenna configurations collaboratively. The project team will design novel methods which enable the optimization of advanced MIMO beam solutions specifically tailored to the highly diverse and dynamic devices