Cooperative Formation Control Of Unmanned Aerial Vehicles Based On Distributed Model Predictive Control

Unmanned aerial vehicle(UAV)is flexible,safe,and reliable,and has the characteristics of no casualties,which can adapt to complex and extreme environments to complete complex and difficult tasks.However,as people’s requirements for tasks are gradually improved and the completion process of the task is constantly complicated,it is difficult for a single UAV to meet the requirements.Therefore,multi-UAV formation control has become a hot issue in the field of UAV cooperative control.However,as the size of the UAVs continues to expand,the communication burden and the computational complexity of the onboard controller have increased,resulting in a decline in real-time performance of the system.Moreover,in the actual flight process,the problem of data loss of the communication between the UAVs may occur,which leads to untimely information exchange between the UAVs and affects the normal operation of the UAV formation.In order to guarantee the stability and reliability of the UAV formation system and complete the established tasks with high quality,it is of great significance to study the problem of UAV formation.To this end,this project focuses on the formation of UAVs with data loss,and the following aspects are carried out respectively:Firstly,considering the possible single random communication data loss,UAV controllers based on distributed model predictive control are designed for the tracking performance and formation maintenance performance of the UAV formation in this project.Considering the leader-follower formation method,and the different roles and different communication methods of the leaders and followers,the optimization problems that the controllers need to solve are designed respectively.In addition,based on Lyapunov theory,a terminal feedback controller is designed,which utilizes the state feedback to predict the information of the lost states,and guarantees that the assumed predictive states converge to the expected states.The assumed predictive states are applied to the optimization problem in the controller to keep the UAV formation in a predetermined formation.Meanwhile,considering the error between the assumed predictive states and the actual predictive states,the collision avoidance constraint and the error constraint are introduced to ensure the safety and accuracy of the formation.Then,for the reason that the consecutive data loss will cause a greater threat to the control performance of UAVs,considering the consecutive random communication data loss problem in the UAV formation,the terminal feedback controller is designed by selecting the appropriate Lyapunov function.The state feedback method is used to estimate the future states according to the optimization states of the previous multiple moments,so that the assumed predictive states are as close as possible to the actual states.The assumed predictive states are applied to the distributed model predictive controller to solve the optimization problem.The collision avoidance constraint and the error constraint are also considered for the stability of the UAV formation when the problem of consecutive data loss occurs.Finally,considering the computation burden caused by the computation of controller in the actual formation task,the event-triggered mechanism is utilized to complete the formation control of UAVs.According to the possible situation of the flight process of UAV formation,and fully considering the stability of formation system,the characteristics of the controller and the possibility of data loss,three event-triggered conditions are designed.Based on the distributed model predictive control,the controller only needs to solve the optimization problem at the moment when the event-triggered conditions are met,which cuts down the consumption of the communication greatly and reduces the amount of calculation.