Research On Task-Driven Intelligent Routing Algorithm For UAV Ad Hoc Network

Unmanned aerial vehicles(UAVs)have emerged as a cost-effective solution for a variety of applications,including forest fire detection and post-disaster relief operations,due to their high flexibility and resilience.These applications require the acquisition of large amounts of data through sensors,and end-to-end data transmission with distinct priorities.However,traditional mobile ad hoc network routing algorithms have been primarily focused on maximizing network throughput or minimizing transmission delay,with limited attention paid to differentiated designs that cater to diverse service requirements.To address this gap,this article presents a novel approach for developing task-driven distributed intelligent routing algorithms and energy-efficient routing strategies for large-scale perception task execution of multiple UAV systems.These algorithms take into account the various transmission urgency and energy constraints of UAVs,and provide distributed intelligent routing and energy-efficient routing strategies for various transmission priorities.The specific research work is explicated as follows:Firstly,this thesis proposes an actor-critic algorithm framework-based task-driven distributed intelligent routing algorithm(A2C-Routing)that ensures real-time delivery of urgent transmission tasks while maximizing the delivery rate of non-urgent tasks.Each UAV acts as an independent intelligent agent,utilizing the actor-critic algorithm for learning and updating routing strategies.The critics of neighboring intelligent agents conduct statistical learning of end-to-end transmission delay through local interaction.The actor updates the routing strategy in real-time based on the network status and task requirements feedback provided by the critic.To handle dynamic topology changes,invalid state masking and invalid action masking mechanisms are designed to enable rapid routing strategy updates in dynamic environments.Simulation results demonstrate that the A2C-Routing algorithm outperforms traditional routing algorithms by achieving routing strategies based on different transmission business priorities in response to changes in network status.Secondly,this thesis proposes a joint decision-making framework for data fusion and energy-efficient routing in a heterogeneous UAV network under task-driven distributed routing scenarios.Based on this framework,a distributed data fusion and energy-efficient routing intelligent joint decision-making algorithm(DFER)is designed.The algorithm employs a double actor-critic algorithm(DAC)that enables each intelligent agent to perform distributed data fusion and energy-efficient routing joint decision-making.To avoid conflicts during the distributed fusion decision-making process,this thesis designs a distributed fusion mode negotiation mechanism,where each intelligent agent combines network status to negotiate fusion modes in order to form consistent data fusion decisions.Simulation results indicate that the DFER algorithm significantly reduces transmission energy consumption and extends network lifetime compared to the A2C-Routing algorithm.