A Multi-Level ML-Based Optimization Framework for IIoT Networks with Distributed IRS Assisted UAVs.

The development of the fifth generation (5G) of cellular systems enables the realization of densely connected, seamlessly integrated, and heterogeneous device networks. While 5G systems were developed to support the Internet of Everything (IoE) paradigm of communication, their mass-scale implementations have excessive capital deployment costs and severely detrimental environmental impacts. Hence, these systems are not feasibly scalable for the envisioned real-time, high-rate, high-reliability, and low-latency requirements of connected consumer, commercial, industrial, healthcare, and environmental processes of the IoE network. The IoE vision is expected to support 30 billion devices by 2030, hence, green communication architectures are critical for the development of next-generation wireless systems. In this context, intelligent reflecting surfaces (IRS) have emerged as a promising disruptive technological advancement that can adjust wireless environments in an energy-efficient manner. This work utilizes and analyzes a multi-node distributed IRS-assisted system in variable channel conditions and resource availability. We then employ machine learning and optimization algorithms for efficient resource allocation and system design of a distributed IRS-enabled industrial Internet of Things (IoT) network. The results show that the proposed data-driven solution is a promising optimization architecture for high-rate, next-generation IoE applications.