In order to study the autonomous control problem of unmanned aerial vehicle(UAV)networks in emergency communication scenarios,multiple rotary-wing UAVs equipped with 6G base stations were used,and a low-altitude and single-layer UAV network was formed through the interconnection between UAVs,thus providing wireless network services to users in the ground task area.A numerical model was constructed for typical scenarios,and a multi-agent reinforcement learning method was used to solve the optimization control strategy of the UAV network.The effects of the number of UAV base stations and the communication distance of UAV base stations on the wireless network coverage in the task area were investigated.The research results indicate that the optimization control strategy obtained by using reinforcement learning methods can converge well.The learning curves of the UAV network coverage score and fairness coverage index have similar trends.The curves rapidly increase between the 1000th and 2000th episode and then change slowly.Under the conditions of a communication coverage distance of 1 km and a flight altitude of 300 m for UAV base stations,the UAV network coverage score increases by 53.28％,and the fairness coverage index increases by 43.57％when the number of UAV base stations increases from 3 to 7.Under the condition of five UAV base stations and a flight altitude of 300 m,the UAV network coverage score increases by 86.01％,and the fairness coverage index increases by 41.47％when the communication distance of the base stations increases from 1.0 km to 2.5 km.
维普
万方数据