UAV Trajectory Planning Method For Capacity Coverage Balance

In recent years,with the rapid development of unmanned aerial vehicle(UAV)technology,UAV communication has gradually become a new solution in the field of emergency communication.UAVs have controllable mobility and strong environmental adaptability,and can provide reliable communication services for rescue personnel and trapped individuals in disaster areas by carrying communication equipment.During the UAV’s service period,effective trajectory planning can be used to reduce communication costs and improve communication quality.Therefore,UAV trajectory planning is an important research direction in the field of UAV communication.Currently,research on UAV trajectory planning mainly focuses on stationary ground user scenarios,aiming at a single capacity or coverage indicator,which has certain limitations.To address this problem,this thesis first studied a single UAV trajectory planning method for capacity and coverage balance.Considering that a single UAV cannot effectively serve all users when the number of users increases,this thesis further extended the single UAV to multiple UAVs and studied the trajectory planning method for capacity and coverage balance for multiple UAVs.The main innovations and conclusions of this thesis are summarized as follows:(1)For the UAV trajectory planning problem in a single UAV scenario,this thesis proposes two solutions.Solution one is a single UAV trajectory planning scheme based on artificial potential field theory.In this scheme,a virtual force function is established based on the system’s capacity and coverage rate indicators,and the UAV’s position is guided by real-time virtual force size and step length to achieve real-time trajectory planning of the UAV communication platform.Through simulation analysis,this scheme can enable the UAV to stably follow the user and provide com munication performance close to that of the static deployment scheme.Solution two considers long-term benefits and user mobility prediction,and establishes the studied problem as a Markov decision process with uniform time steps,designing a trajectory planning algorithm based on deep deterministic policy gradients.Through simulation comparison with the benchmark algorithm,the proposed algorithm has better capacity and coverage performance.(2)For the UAV cooperative trajectory planning problem in a multiple UAV scenario,this thesis proposes a solution based on multi-agent reinforcement learning.First,the cooperative trajectory planning problem for multiple UAVs to provide communication for multiple ground mobile users is modeled.Considering that the information observed by multiple UAVs is mutually independent,the optimization problem is further modeled using a stochastic game model,and a trajectory planning scheme based on distributed deterministic policy gradients is designed to solve the problem.Through simulation comparison with centralized reinforcement learning-based trajectory planning algorithms and benchmark algorithms,the proposed algorithm can ensure capacity and coverage balance of the entire system during service,and its capacity and coverage performance are better than that of the benchmark algorithm and close to the centralized algorithm.