Deep Learning Techniques for OFDM Systems
M. Meenalakshmi, Saurabh Chaturvedi, Vivek K. Dwivedi
Abstract
Orthogonal frequency division multiplexing (OFDM) is a popular multicarrier technique in communication system owing to its robustness against multipath fading and less complexity. Deep learning (DL) approach is more accurate and efficient than traditional approaches, which seeks more attention not only in the fields of natural language processing, video processing, speech and audio processing but also in the field of communication systems. The application of DL technique in OFDM systems supports better system performance, peak-to-average power ratio (PAPR) reduction, and improvement in spectral efficiency. This paper presents a detailed review of recent developments in DL techniques for OFDM systems to improve its performance in terms of bit error rate, signal-to-noise ratio, and PAPR. Various DL frameworks available for architectural design and processing are also explained in this article. The paper is concluded with a discussion on research gaps and challenges for future investigation and development.