Research In UAV Network Energy Saving Approaches For Ground Multicast Service Transmission

With the rapid development of unmanned aerial vehicle(UAV)communication technology,UAV has been widely used in military reconnaissance,environmental detection,film and television aerial photography,post-disaster rescue and other fields.However,due to the lack of UAV's endurance,the further development of UAV communication is severely limited.Therefore,this thesis focuses on the energy consumption of UAV,and makes an in-depth study on the energy conservation approaches of UAV network,to minimize the energy consumption of UAV for multicast transmission.The main research contents and contributions are summarized as follows.First,considering the specific scenario of UAV network multicast transmission,a common information(Cl)file is transmitted to all ground terminals(GTs)by UAV multicast.To reduce the UAV energy consumption,an optimization framework of machine learning empowered joint multicast grouping and UAV trajectory design is proposed.First,the framework assumes that multicast groups are determined,a centroid-adjustable traveling salesman problem(CA-TSP)inspired iterative optimization algorithm is proposed for UAV flight trajectory design.Second,a compressed-feature regression and clustering machine learning(C2ML)algorithm is proposed to compress the GTs' location distribution into a one-dimensional silhouette coefficient,and the support vector regression(S VR)model is trained to determine the number of groups that guides K-means clustering to approach the optimal multicast grouping.Finally,the framework integrates the offline C2ML multicast group module and the online CA-TSP trajectory optimization module.The simulation results show that the proposed scheme improves the performance by 24.7%compared with the comparison scheme,and significantly reduces the total energy consumption of UAV.Next,considering the general scenario of UAV network multicast transmission,using UAV networks to transmit CI files for more GTs,and taking the mobility of GTs into consideration,an optimization framework of machine learning empowered joint multicast grouping and UAV networks scheduling is proposed.First,assume that multicast groups are determined,the optimization problem is transformed into minimum circle coverage problem and multiple vehicles scheduling problem.Then,an UAV network scheduling optimization algorithm is proposed to determine the UAVs'trajectory.Next,adopt the Gaussian process(GP)to predict GTs'location and then adjust UAVs' location.Next,the C2ML module is also used to determine the number of groups that guides K-means clustering to approach the optimal multicast grouping.Finally,the framework integrates the offline C2ML multicast grouping module and the online UAV networks scheduling module.The simulation results show that the proposed scheme improves the performance by more than 33%compared with the comparison scheme,and significantly reduces the total energy consumption of UAV networks.