SPARSE LINEARIZATION AND HARDWARE FOR 5G AND BEYOND

Purpose and goal:

In this project, we have studied linearization of power amplifiers in base stations, focusing on how to exploit the fact that the signal is typically sparse in frequency (only a small part of the bandwidth is used at each time) and spatially/antenna (a few users are served with many antennas). The results clearly show that by exploiting sparsity we can reduce the complexity, and thus the energy consumption, in both analog and digital hardware.

Expected results and effects:

The project has resulted in 7 publications, of which 4 are journal articles. In addition, a patent has been awarde on an algorithm based on project results. In addition, two test beds have been constructed, one at Chalmers and one at Eriksson.

The expected effects are that knowledge and results from these project will over time be implemented in commercial products, with significantly more energy efficient base stations as a result. Approach and implementation:

The work was divided into work packages; one that focused on hardware, another on algorithms, and a third on testbed implementation. Generally, the setup has worked well, possibly with the exception that testbeds took a long time to get up and running, and little time at the end of the project was left to actually implement the algorithms in a testbed. Each work package had its own meetings, and at regular intervals we had joint meetings across the project to synchronize the results.