Cross-Layer Interference Management: Bringing Interference Alignment to Reality
Planning to Explore via Self-Supervised World Models
Cross-Layer Interference Management: Bringing Interference Alignment to Reality

Agrim Gupta
agg003@ucsd.edu
Sajjad Nassirpour
sajjad.nassirpour@ucdenver.edu
Dinesh Bharadia
dineshb@ucsd.edu
Alireza Vahid
alireza.vahid@ucdenver.edu


In today's wireless networks, the typical operating paradigm is to have different users occupying different frequency band such that they don't interfere. Multiple user MIMO (mu-MIMO) has been previously attempted to solve this problem and have multiple users communicating over a single frequency band. However, mu-MIMO implementations have never been robust since it requires favorable channel conditions not always guaranteed in real wireless channels. Typical reasons for this is that user locations being close to each other (Fig. 1) makes the channels correlated, and does not allow for efficient interference separations, as interference is almost indistinguishable from the intended users' signals. In this project, we aim to make the mu-MIMO gains robust by creating favorable channel conditions to enable multi-user communications. This is done by implementing interference alignment (IA), by having a larger number of antennas at the base station, which can be turned on-off. The goal of the project is to create variations in the received wireless channel by turning these antennas on-off strategically, in a way that the interfered channel always aligns in a single direction (Fig. 2), wheres the targeted user channel remains variable. This guarantees a good operating point for interference management, since the interference is aligned across the switching antenna states and thus can be suppressed easily.


Website Template Originally made by Phillip Isola and Richard Zhang for colorful ECCV project; the code can be found here.