Machine Learning-Based Non-Orthogonal Frequency Division Multiplexing for Data Overlay Transmission and Reception

Abstract

The rapid evolution of mobile communication technologies like 5G and the anticipated emergence of 6G have necessitated enhanced resource utilization and reception efficiency in wireless communication systems. One of the challenges faced in existing systems is the resource competition arising from the overlapped transmission of pilot frequency and data. This paper addresses this issue by proposing a deep learning-based solution that leverages machine learning techniques to decode transmitted data efficiently. Through a combination of non-orthogonal pilot-frequency and data-overlapped transmission schemes, the proposed solution effectively mitigates interference problems, leading to improved resource utilization and reception performance. Experimental and simulation analyses validate the efficacy and feasibility of the proposed approach, showcasing a decoding accuracy of up to 93% under specific conditions.