Machine Learning-Assisted Interference Management in the 6G UAV Networks with Soft Frequency Reuse

Sixth-generation (6G) wireless networks are envisioned to be ubiquitous. To ensure ubiquity and seamless connectivity, networks with unmanned aerial vehicles (UAVs) are envisioned as a potential solution. However, many challenges are still ahead to make such networks practically feasible. Recently, soft frequency reuse (SFR) is exploited for reducing the interference of UAV-assisted networks. However, the existing deep learning-based solution in allocating resource plans among the UAVs in a multi-UAV network suffers from low accuracy. In this chapter, we propose a machine learning model to address the low-accuracy issue of UAV-assisted networks with SFR. We use nine classifiers to allocate the resource plans. It is found that the random forest provides the highest accuracy. It outperforms the deep learning-based solutions by up to 7.3% more accuracy. We also investigate to find an answer to the question, “Should we consider the positions of all UAVs to allocate a resource plan to a specific UAV?” Through simulations, it is also found that the positions of all UAVs are not required to be considered in allocating resource plans in a network with 12 UAVs. However, the positions of all UAVs should be considered for the network with a smaller number of UAVs.