Alternating optimized RIS-assisted noma and nonlinear partial differential deep reinforced satellite communication

Abstract

Satellite communication is crucial for 6 G industrial wireless networks due to its comprehensive communication services, but traditional methods face challenges such as high latency, limited bandwidth, high cost, and integration with terrestrial networks. This highlights the need for innovative communication methods to boost satellite communication efficiency, reliability, and coverage within the context of 6 G industrial applications, which is attempted to address by this research. RISs (Reconfigurable Intelligent Surfaces) are thin sheets with tiny antennas that are programmed to manipulate radio signals. The RIS has the potential to revolutionize industrial wireless communication by providing greater dynamic and goal-oriented control over radio signals. From there, the wireless environment is being transformed into a service. Besides, the NOMA (Non-Orthogonal Multiple Access) mechanism is a highly effective radio access method in advanced wireless communications that optimizes spectrum utilization and offers improved spectrum efficiency, lower latency, more reliability, and increased connectivity. The current industrial communication network faces limitations such as limited capacity, signal interference overload, complex and expensive infrastructure, and latency issues. The fusion of RIS and NOMA is seen as promising strategy to meet the stringent demands of next-generation networks for immense connectivity. It is particularly innovative for the demanding environment of smart factories by enabling efficient resource sharing, dynamic signal control, ultra-reliable, low-latency communication, cost-effective network, overcoming the signal interference and simplifying the network infrastructure. In this article, we propose a novel approach for device-to-device satellite communication in industrial networks. This approach leverages a RIS (Reconfigurable Intelligent Surface) -assisted cooperative network with NOMA (Non-Orthogonal Multiple Access) and NL-PDDRL (Nonlinear Partial Differential Deep Reinforced Learning) termed AONPDDR (Alternating Optimized and Nonlinear Partial Differential Deep Reinforced) communication. This article focuses on important aspects related to industrial communication networks such as power consumption, energy efficiency, data delivery rate, data loss rate, throughput, latency, coverage, false alarm rate, and bit error rate. The Alternating Optimization-based RIS configuration algorithm optimizes reflecting elements in RIS by maximizing channel vectors and providing a sophisticated upper bound on the highest number of reflective elements. Next, The NL-PDDRL algorithm is designed to function effectively despite environmental uncertainties and constraints. The results indicate that the proposed AONPDDR method outperforms previous benchmark methods in system-achievable aspects.

Keywords

• Reconfigurable intelligent surfaces
• Non-orthogonal multiple access
• Alternating optimization
• Nonlinear Partial Differential
• Deep Reinforced Learning