LAUSR

Adaptive routing and efficient packet delivery in Flying Ad Hoc Networks (FANETs) are significant challenges due to underlying environment constraints, such as dynamic nature, mobility, and limited connectivity. With the increasing number of machine learning (ML) applications in wireless networks, FANETs can benefit from these data-driven predictions. This letter proposes a Packet Arrival Prediction (PAP) routing protocol to improve transmission link reliability. Primarily, we apply a Long Short-Term Memory (LSTM) model to predict the packet arrival of each UAV, seeking to avoid the high-traffic UAVs, which cause packet loss significantly. Then, we formulate the routing decision issue as an optimization problem, which attempts to find an appropriate path by a proposed constrained sorting approach, in order to make joint yet fast routing decisions. The simulation results demonstrate that the PAP routing protocol outperforms the existing manifold protocols in the aspects of Packet Delivery Ratio (PDR) and delay.