Dynamic collaborative optimization coverage algorithm of UAV based on trajectory prediction in Internet of Vehicles

This research presents a dynamic pre-deployment strategy based on vehicle trajectory prediction information to address the issues of coverage gaps of base stations and local traffic congestion in urban vehicle networking. Under the architecture of distributed federated learning and blockchain, numerous UAVs equipped with edge computing servers eliminate the central aggregation node and use an enhanced Raft method to train a unified Seq2Seq-GRU trajectory prediction model. In the training cycle, the scheme selects the nodes to perform parameter aggregation and model update based on the quantity of data given. Second, based on the model's prediction findings, this study offers an enhanced virtual force guide deployment method that directs the UAV to dynamically deploy across each virtual force in order to increase the vehicle's access rate and communication quality.