Towards Fully DL-Driven RF: A Systematic Survey of Deep Learning for Wireless Transceiver Signal Processing

The traditional algorithm-based signal processing techniques have shown to be ineffective in real-world implementations with the increased demand on modern wireless communication systems stemming from the rapidly growing Internet of Things (IoT) ecosystems and industrial automation. Conventional Digital Signal Processing (DSP) techniques have shown limited adaptability to adverse channel conditions and algorithmic complexities, therefore diminishing their effectiveness in real-world deployments. More recently, there has been a growing interest in utilizing Deep Learning (DL) as an alternative methodology that has shown to be comparable, and in some cases superior, to the traditional methods in low Signal-to-Noise Ratio (SNR) and high mobility and non ideal channel situations. This particular area of research

aims to systematically study the variety of ways that DL has been applied in the development wireless transmission methods, specifically in this case, focusing on the advancements of the technology based on Orthogonal Frequency Division Multiplexing (OFDM) which serves as the backbone for the various iterations of Wi-Fi, 4G LTE, 5G and Digital Video Broadcasting (DVB). The intended outcome is an evaluation and analysis, through simulation, of the latest advances of DL methodologies applied to the OFDM receiver structure, evaluating their ability to address the traditional limitations of DSP and providing a pathway for the development of radio frequency systems that are fully DL based.

Message: Hi FE, I need help with “Towards Fully DL-Driven RF: A Systematic Survey of Deep Learning for Wireless Transceiver Signal Processing” in “Electronics & Communication Engineering”