Selective Early Retransmissions Based on Channel Classification for UAV Control Link

Control data for unmanned aerial vehicles (UAVs) are characterized by stringent reliability and latency requirements, thus transmission parameters must be suitably set. In particular, packet retransmission can significantly improve the reliability but risks compromising latency requirements. In this paper an efficient proactive retransmission scheme based on channel classification is proposed. The retransmission occurs without waiting the receiver feedback, and is selectively used for a subset of packets. Differently, classical proactive schemes foresee the retransmission of all packets with a consequent waste of resources. The selection is based on the classification of the channel behavior before the next channel state report and on the aging of the channel state information. A deep recursive neural network is used to classify the UAV channels. Numerical results show that the proposed approach allows to reach a suitable tradeoff between reliability and spectral efficiency compared with different solutions.