A Research Of Multi-UAV Real-time Task Allocation

Unmanned aerial vehicle(UAV)is the inevitable outcome of the development of modern science and technology.As the rapid development of science and technology,especially artificial intelligence and automation,the automation degree of UAV system has been improved continually,and the related studies have drawn increasingly more attention of researchers.Currently,aspects of flight path programming,task allocation,flight detection and obstacle avoidance,cooperative engagement,etc.are covered within the research of academic in home and abroad.This thesis focuses on multi-UAV real-time task allocation technology,which takes along different type of resource.The UAV task assignment is defined according to the environmental information provided by the satellite monitoring equipment,the control and deployment system,particular function and parameters to carry out the deployment of the UAV's missions,and can be adjusted according to the real-time environment information,the objective is to provide a task allocation scheme with high safety and high penetration probability,to ensure the best overall combat effectiveness of the UAV system.In this thesis,the task assignment of multi-UAV is divided into two stages.The first stage is the task pre-allocate based on the global information before task execution.The second stage is the dynamic adjustment of the decision scheme according to the changes of the real-time environment in the process of the task execution,and the coordination of the conflict problems.The main contents of this thesis include the following aspects:1.Communication structure and task allocation algorithms for UAV system are studied,and the feasibility and stability of the existing algorithm in the complex battlefield environment described in this thesis are analyzed.To find a suitable algorithm,the task assignment problem is modeled and divided into two stages,which say the task pre-allocate and dynamic real-time allocation,the problem of each stage is solved with a specified algorithm.2.In the process of the task pre-allocate,we consider the scale of UAV system,the complexity of battlefield environment and other factors.To solve the global optimality of pre-allocate,an approach based on the Cross-Entropy method based on probability estimation is used in this thesis,and the determination of the allocation scheme is speed up by the application of column encoding and introduction of local search strategy.3.In the process of real-time task allocation,the pre-allocate scheme is dynamically adjusted according to the unpredictable situation in the real-time environment,the contract net protocol(CNP)based on market mechanism is applied,which by simulates market economic activities to implement dynamic adjustment of UAV in this thesis.Furthermore,this thesis selects the suitable strategy to make perfect the mechanism of the scope of tender,the number of bids,the duties of the tenderer and coordinate the possible conflicts.