Training, Prediction and Adaptation Using AI for Communication Channel “Ground Control Station – Satellite - Aerial Relay Drone – UAV”

Article demonstrates a way for adaptation Unmanned Aerial Vehicle (UAV) communication channel using Artificial Intelligence (AI). Simulation was provided on the base of original model for communication channel “Ground Control Station – Satellite – Aerial Relay Drone – UAV”, which was built using NetCracker software. Impact of bit error rate and packet size on latency, impact of data rate and packet size on latency, impact of bit error rate and packet loss probability on satellite, and impact of bit error rate and packet size on channel throughput were studied. For adapting parameters depending on the delay and the number of bit errors, a linear regression model was created and trained to predict the size of the transmitted packet based on the values of the delay and the level of bit errors. Adaptive transmission was implemented based on the prediction of the packet size for a given delay and dynamic change in the packet size.