Deep Learning Neural Receiver for Organic Communication Channels

Organic communication channels (OCCs) are any hydrocarbon-based media used to communicate data. OCCs have gained attention as an alternative to short-range RF communication, offering advantages such as higher data rates, improved spectral efficiency, and greater power efficiency. OCCs show tremendous potential when applied to smart farming and smart health. To date, research done on OCCs are limited to human body and soil-based communication. Relatively high error rates in existing OCC PHYs gave way to the exploration of other PHY configurations. Consequently, this paper develops a neural receiver architecture that is IEEE 802.15.6 standard compliant for the purpose of OCC communication using soil and vegetation-based OCCs. The results show this receiver, when integrated with an IEEE 802.15.6 standard compliant transmitter, having generally a better error performance compared to what currently exists in the literature, as well as the one designed using conventional decoding techniques, even under different channel distances. The relatively low error rates make it possible for voice and video communication in OCCs.