Machine Learning-Driven Performance Analysis of Compressed Communication in Aerial-RIS Networks for Future 6G Networks

Muhammad Farhan Khan, Muhammad Ahmed Mohsin, Zeeshan Alam, Muhammad Saad, Muhammad Waqar · 2025

Read Machine Learning-Driven Performance Analysis of Compressed Communication in Aerial-RIS Networks for Future 6G Networks by Muhammad Farhan Khan, Muhammad Ahmed Mohsin, Zeeshan Alam, Muhammad Saad, Muhammad Waqar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-01911.txt.

Introduction

Muhammad Farhan Khan and four co-authors design a wireless system meant to raise data rates in crowded future 6G cities. Their model combines drone-carried reconfigurable intelligent surfaces, non-orthogonal multiple access, and coordinated transmission between cells. Each piece attacks a different limit: capacity, spectrum use, and interference.

A practical snag is feedback: receivers must send quantized phase settings back to the surface, but the channel cannot carry much. The authors use a machine learning autoencoder to compress that feedback while keeping accuracy high. They also study how the number of surface elements and power sharing between users affect each person's data rate.

Catalog record

CREATORMuhammad Farhan Khan, Muhammad Ahmed Mohsin, Zeeshan Alam, Muhammad Saad, Muhammad Waqar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-01911

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Machine Learning-Driven Performance Analysis of Compressed Communication in Aerial-RIS Networks for Future 6G Networks Muhammad Farhan Khan, Muhammad Ahmed Mohsin, Zeeshan Alam, Muhammad Saad, Muhammad Waqar Abstract In the future 6G and wireless networks, particularly in dense urban environments, bandwidth exhaustion and limited capacity pose significant challenges to enhancing data rates. We introduce a novel system model designed to improve the data rate of users in next-generation multi-cell networks by integrating Unmanned Aerial Vehicle (UAV)-Assisted Reconfigurable Intelligent Surfaces (RIS), Non-Orthogonal Multiple Access (NOMA), and Coordinated Multipoint Transmission (CoMP). Optimally deploying Aerial RIS for higher data rates, employing NOMA to improve spectral efficiency, and utilizing CoMP to mitigate inter-cell interference (ICI), we significantly enhance the overall system capacity and sum rate. Furthermore, we address the challenge of feedback overhead associated with Quantized Phase Shifts (QPS) from the receiver to RIS. The feedback channel is band-limited and cannot support a large overhead of QPS for uplink communication. To ensure seamless transmission, we propose a Machine Learning Autoencoder technique for a compressed communication of QPS from the receiver to RIS, while maintaining high accuracy. Additionally, we investigate the impact of the number of Aerial RIS elements and power allocation ratio for NOMA on the individual data rate of users. Our simulation results demonstrate substantial improvements in spectral efficiency, outage probability, and bandwidth utilization, highlighting the potential of the proposed architecture to enhance network performance. † † aftertitle: I Introduction Reconfigurable Intelligent Surfaces (RIS) have garnered significant attention as a solution to enhance the performance of wireless communication networks beyond the fifth generation (B5G) and sixth generation (6G). As urbanization and skyscrapers increase in cities, there is a growing need for an efficient and cost-effective method to enable Non-Line-of-Sight (NLOS) communication, which can be facilitated by RIS. A RIS consists of an array of passive reflective elements on a 2-D meta-surface. Each element can be individually programmed to enhance signal coverage by adjusting both phase and amplitude, enabling efficient and intelligent steering so that the signals are constructively combined at the receiver’s end, resulting in higher data rates, improved spectral efficiency, and ubiquitous connectivity while being eco-friendly and cost-efficient. RIS creates a Virtual Line of Sight (VLOS) while passively recycling the signal without any external interference. This makes RIS a crucial enabler for the envisioned smart radio environments of 6G networks. [ 1 , 2 ] . The advancement of technology has shifted the surge of bandwidth utilization from Orthogonal Multiple Access (OMA) to Non-Orthogonal Multiple Access (NOMA). OMA, which includes TDMA, FDMA, and CDMA, technologies were susceptible to bandwidth exhaustion that could be utilized for other cell users. In contrast, NOMA (Non-Orthogonal Multiple Access) improves bandwidth efficiency and overall network capacity by allowing multiple users to share the same frequency band simultaneously. This requires successive interference cancellation (SIC) to separate user signals based on power levels [ 3 ] . [ 4 ] , [ 5 ] . By leveraging power domain multiplexing, NOMA can serve multiple users, enhancing throughput and reducing latency [ 4 ] . The integration of RIS with NOMA systems has been shown to further amplify these advantages by optimizing the signal-to-interference-plus-noise ratio (SINR) and providing more flexible resource allocation [ 6 ] . Integrating CoMP with NOMA techniques is promising for mitigating severe ICI effects in multi-cell NOMA networks and improving spectral efficiency [ 7 , 8 , 9 , 10 ] . Coordinated multi-point (CoMP) techniques use high-speed front-haul connections and sharing channel state information (CSI) among BSs, that enable coordinated transmissions and enhance overall network performance. However, the challenge of coordination among all BSs still arises due to issues with CSI inaccuracies, synchronization across cells, and increased signal processing requirements. While the integration of RIS and NOMA has shown promising results in improving spectral efficiency and reducing outage probability [ 11 , 12 ] , the combination of RIS, NOMA, and CoMP has not been thoroughly investigated, especially with aerial RIS deployments. RIS, mounted on platforms such as unmanned aerial vehicles (UAVs), can provide flexible and dynamic coverage enhancements, adapting to changing network conditions in real-time [ 13 ] . According to the author’s best knowledge, no performance analysis exists for Aerial-RIS-aided CoMP-NOMA cellular environments [ 14 , 15 ] . Managing RIS-assisted systems requires phase control for enhanced performance. Many studies have been made for finding optimal phase shifts for different optimization objectives [ 16 , 17 , 18 ] . Previous studies assumed that these phase shifts are already available at the RIS end in quantized form, but that is not the case. These Quantized Phase Shifts require a feedback channel from the receiver to RIS, which is difficult to obtain due to bandwidth limitations. Bandwidth efficiency or Spectrum efficiency is a crucial requirement for efficient data transmission. A promising approach is to use machine learning-based autoencoders for data compression to achieve this efficiency. Autoencoder is a special type of neural network that can learn to represent the data in a lower dimension and then decode it back to its original form, saving significant bandwidth. The paper investigates using autoencoders to compress the QPS symbols into lower dimensions for a smooth transmission over the feedback channel and decode them at the user end. We can achieve substantial bandwidth reduction while maintaining data integrity by leveraging autoencoders’ encoding and decoding capabilities [ 19 ] . Motivated by the literature gap, we present a detailed Machine Learning performance analysis for compressed communication of a system integrating aerial RIS, CoMP, and NOMA. This integrated approach not only enhances performance metrics such as throughput, latency, and coverage but also paves the way for innovative applications in diverse fields ranging from smart cities to autonomous driving [ 20 , 21 , 22 ] . Our contributions are threefold: • We developed a novel system model for aerial RIS-assisted CoMP-NOMA networks, considering strategic placement of RIS for far-user gain enhancement. • We analyze the impact of various parameters, including RIS elements, and power allocation coefficients, on the system’s spectral efficiency and outage probability. • Our system model proposes the far user as a part of two NOMA clusters and our simulation results demonstrate an increase in data rate, network capacity, coverage, and a decrease in outage probabilities. • We compared different autoencoder architectures utilized for compressed communication of encoded QPS bits and observed the reconstruction accuracy of our data from all models. Fig. 1: System Model: Power Allocation for UEs (UEn, UEf) with Reconfigurable Intelligent Surface (RIS) Aided Communication II System Model and Problem Formulation II-A System Description As described in Fig. 2 , we consider a two-cell narrow band CoMP-NOMA system with an Aerial-RIS placed strategically at the cell edge of the two cells to facilitate both cell users. The downlink communication system operates with frequency-flat channels. The Aerial RIS consists of M elements. Let the index for the base stations be \mathcal{C}=\{1,2,\dots,C\} , \mathcal{N}=\{1,2,\dots,N\} for the two near users and \mathcal{F}=\{1,2,\dots,F\} for the far cell user. Moreover, let \mathcal{U}=\{\mathcal{N}\,\cup\,\mathcal{F}\} be the set of combined users in the system. In Fig. 1 a conceptual wireless communication network for a 3D urban environment is visualized, demonstrating the practical use case of our system model. In this urban environment, the direct

Read the complete text