UAV Target Roundup Strategy Based On Biological Swarm Intelligence Behavior

With the continuous development of modern science and technology,unmanned aerial vehicle(UAV)has performance advantages such as small size,high flexibility,and good concealment.In recent years,in the Nagorno-Karabakh conflict and the Russia-Ukraine conflict,UAVs played a crucial role in electronic countermeasures and fire strikes,and had a great impact on the direction of the war situation.UAV cluster warfare has higher efficiency,stronger stability,and is more suitable for executing tasks in complex battlefield environments.It is a new modern combat mode,and many countries have invested a large amount of resources into research in related fields.Collaborative roundup is one of the typical scenarios of UAV cluster operations,which refers to the mutual cooperation of UAVs in the cluster to achieve encirclement control of the target.This paper focuses on the collaborative encirclement problem in UAV cluster operations,and the main research contents are as follows:Firstly,an UAV cluster target rounding strategy based on biological swarm intelligence behavior is proposed.During the operation of UAV cluster,due to the complexity of battlefield environment,each UAV often has certain limitations in obtaining the surrounding environment information and communicating with other units.Therefore,it is of great significance to learn from the distributed and self-organizing characteristics of the intelligent behavior of biological clusters in nature and apply them to the study of cooperative control of clusters.This paper studies and analyzes the movement behaviors of some typical biological groups in nature,and conducts mathematical modeling on the interaction behaviors between individuals in the group and other individuals,as well as the interaction behaviors between individuals and the local environment around them.Combined with the actual characteristics of UAV cluster combat missions,this paper applies them to the target rounding strategy.Secondly,multi-agent deep reinforcement learning is applied to the target rounding problem of UAV cluster.According to the characteristics of the intelligent behavior of biological groups,a reward function in the Multi-Agent Deep Deterministic Policy Gradient(MADDPG)algorithm is designed to guide the UAV to circle the target.In order to solve the problem that the algorithm cannot effectively utilize important sample data,a preferential experience replay mechanism is introduced in the sampling process to improve the learning efficiency.Finally,the proposed algorithm is verified by simulation experiment.The rounding strategies proposed in this paper are all based on the swarm intelligence behavior of organisms.The UAV in the cluster does not need to obtain the global information in the environment,but only needs to perceive the local environmental information around itself,and can successfully control the escape target and complete the rounding task without communicating with each other,which is suitable for the complex environment on the battlefield.Experimental results show that the strategies proposed in this paper have strong robustness and can effectively capture escape targets.