Evaluating the Impact of Clustering Algorithms on the Performance of NOMA-Based UAV Networks

Sammanfattning

Non-orthogonal multiple access (NOMA) integrated with unmanned aerial vehicle (UAV)-assisted networks has received significant attention in recent years due to its potential to significantly transform wireless communication systems, offering enhanced spectral efficiency and coverage flexibility. User clustering in such networks plays a crucial role in optimizing system performance and resource allocation. The integration of UAVs as aerial relays in NOMA networks enables simultaneous service to multiple users through power-domain multiplexing, where efficient user clustering becomes essential for optimal network performance. Various clustering algorithms can be employed to group users based on their spatial distribution and channel conditions. In this thesis, we consider a downlink NOMA-based UAV network where multiple UAVs operate as decode-and-forward (DF) relays in half-duplex (HD) mode. The system employs different clustering algorithms, e.g., K-means, K-means++, K-medians, etc., to organize randomly distributed users into optimal clusters. The communication occurs over two hops—the base station (BS) to UAVs and UAVs to users, both experiencing Nakagami-m fading conditions. Comprehensive simulations are conducted using MATLAB to model and evaluate the performance of the NOMA-based UAV network under various clustering algorithms. The system's performance is assessed through two key metrics, i.e., outage probability and sum rate. The simulation framework considers realistic network conditions, such as fading and path loss, to provide accurate performance evaluations. The results demonstrate that the choice of clustering algorithm significantly impacts system performance, particularly as the number of UAVs increases. Furthermore, the study explores how varying fading environments and UAV transmit signal-to-noise ratios (SNRs) influence network reliability and capacity. These findings underscore the importance of clustering strategies in maximizing the potential of NOMA-based UAV networks for next-generation wireless communication systems.