Temporal deep learning assisted UAV communication channel model for application in EH-MIMO-NOMA set-up

The radio frequency (RF) spectrum is crucial for effective deployment of unmanned aerial vehicle (UAV). The unpredictability of the communication channel restricts link reliability and quality of service (QoS) of the UAV's deployment. Though several approaches have already been reported to model the communication channel of the UAV, the necessity of optimal spectrum utilization, better link reliability and QoS requires that new methods be explored for effective representations of propagation variations and path gains present in the UAV flight profile. Here, we report the design of a learning aided model of the UAV communication channel. We use a deep learning (DL) based method which relies upon different taped delay line (TDL) driven layers of gated functions implanted as part of specifically designed networks for capturing the channel state information (CSI) of the propagation medium. The key part is the use of context processing and recovery layers formed by these TDL driven gated function structures which provide performance enhancements. The proposed model is trained and tested with synthetic and actual data and is supported by energy harvesting attributes. It has been found to be effective in modeling UAV channels while deployed with multi input multi output (MIMO) and non-orthogonal multi access (NOMA) set-up in urban areas and in platforms moving with a maximum velocity of 60 kmph.