Research On Adaptive Cooperative Control Of UAV Swarm Under Constrained Network

With the development of UAV swarm technology,UAV swarm have more and more application scenarios.However,due to the complexity of the real environment,the factors that interfere with the UAV swarm communication network can be seen everywhere.Therefore,a UAV swarm control technology with certain adaptability to the network state is proposed in this thesis,which is very important for UAV swarm control.At present,most UAV swarm control technologies only focus on the control model itself and make simple assumptions about the network communication module.As a result,when confronted with the constrained network,these UAV swarm control technologies cannot provide stable control ability and make adaptive adjustment timely.Therefore,in order to expand the application field of UAV swarm control technology,it is necessary to study the UAV swarm control technology under restricted network to improve the control effect of the control model under the restricted network.In this thesis,the existing work on UAV swarm control and adaptive control is surveyed and analyzed.In order to make UAV swarm control technology has certain adaptive capacity to the constrained network state,we focus on the build of UAV swarm constrained network model and adaptive control of UAV swarm under constrained network.The construction of UAV swarm constrained network model is mainly to model the UAV swarm network communication module,and to build a more reasonable UAV swarm communication mechanism and communication constrained mechanism.Based on the influence of constrained network on UAV swarm control,the adaptive control of UAV Swarm under constrained network proposes a control model that can make UAV swarm control adjust adaptively based on the current flight environment and network conditions,finally improve the cooperative performance of UAV swarm under constrained network.The main results of this research are as follows:(1)Combined with the relevant theories of flying ad hoc network,the communication mechanism of UAV swarm control model is modeled.Meanwhile,communication delay model and communication packet loss model are constructed,and the communication failure mechanism based on link is realized by combining the characteristics of high flying speed and high flexibility of UAV.(2)Based on the constructed constrained network model,combining the existing UAV swarm control model and the adaptive control based on reinforcement learning,we propose a UAV swarm adaptive control model under constrained network,which can realize the adaptive control of UAV swarm under constrained network.Meanwhile,combined with the UAV swarm communication mechanism constructed,a network monitoring index based on the importance of communication link is proposed to calculate the robustness of the network and carry out better adaptive control.(3)By introducing the idea of decompose and then couple,we construct a network structure that can extract features from UAV swarm state information,and the long short term memory network is combined to process the timing information to improve the effectiveness of the adaptive adjustment strategy.(4)By designing multiple UAV swarm simulation scenarios,we verified the validity and reliability of the constrained network model.Meanwhile,the adaptability and superiority of the UAV swarm adaptive control model under constrained network is verified.