Research On Distributed Formation Control Problem Of Multi-quadrotor UAV

With the continuous expansion of the application field and the continuous improvement of performance requirements,it is gradually difficult for a single unmanned aerial vehicle(UAV)to meet the requirements of increasingly complex scenarios.UAVs formation control has become an important direction for the future development of the control field due to the advantages of high task execution efficiency and strong fault tolerance.Due to the characteristics of low autonomy,high dependence on the control station,poor real-time performance,and easy to be affected by external interference,unmodeled dynamics,actuator failure and collision between formation individuals,centralized UAV formation control is difficult to meet the actual needs of UAV formation control.In order to improve the cooperative control ability of UAV formation,it is necessary to study the distributed formation control technology and establish a flexible and efficient formation flight control scheme to solve the above problems.Therefore,this thesis mainly takes the quadrotor UAV as the research object,based on the multi-agent consistency theory,combined with a variety of nonlinear intelligent control algorithms such as backstepping method,adaptive control and neural network control,and proposes several different adaptive distributed cooperative formation control schemes.The main research contents and achievements of this thesis are as follows:Firstly,a distributed formation control strategy based on adaptive neural network is proposed for the position and attitude formation control problem of quadrotor UAV with unmodeled dynamics and external disturbances.Newton-Euler formula is used to describe the dynamic model of quadrotor UAV.In order to deal with the problem easily,the system model is equivalently transformed into position subsystem and attitude subsystem which are easy to analyze.By using algebraic graph theory to describe the communication topology of the multiquadrotor UAV system,the communication mechanism between the UAV and the virtual leader is established,and a distributed formation control scheme based on backstepping method is designed.Each UAV can update its position and speed online according to the state information of adjacent UAVs to form a predefined formation.The designed formation control scheme is combined with the neural network sliding mode control method to deal with the problems of unmodeled dynamics and external disturbances in the quadrotor UAV formation flight control system.Lyapunov stability theory is used to prove that the tracking error of UAV converges to zero in the adjustable neighborhood.In the simulation part,the neural network adaptive formation control method is compared with the neural network sliding mode adaptive formation control method,and the effectiveness of the method is verified.Secondly,while considering the unmodeled dynamics of the system and improving its ability to resist external disturbances,a distributed adaptive neural network fault-tolerant formation control method is proposed for multi-quadrotor UAV systems with actuator faults.By introducing multiplicative and additive fault coefficients,a nonlinear dynamics model of the quadrotor UAVs with variable actuator faults is established.In practical applications,the communication network topology structure of multi-quadrotor UAVs may change due to disturbance or fault,etc.The switching topology structure is used to describe the communication network between multi-quadrotor UAVs,and the complex switching network is converted to the topology structure connected only for each switching subset,which reduces the difficulty of system analysis.By designing a neural network adaptive sliding mode formation controller,the distributed formation control problem of UAVs is solved when the dynamic model of the system is changed due to actuator failure.The uniform final bounded control system is proved by stability analysis.According to the comparison and analysis of simulation results,the proposed adaptive neural network sliding mode fault-tolerant formation control scheme has better control performance and fault-tolerant ability than adaptive formation control scheme,and the proposed control algorithm has better robustness in dealing with unmodeled dynamics and external disturbances.Thirdly,considering the requirements of collision avoidance task during formation transformation,the artificial potential field method is extended to three-dimensional space according to the three-dimensional spatial dynamics model of multi-UAV system,and the problem of collision avoidance control among individuals in multi-UAV formation flying system is studied.The directed spanning tree is used to optimize the topology structure of the UAV communication network,which can save communication bandwidth resources and improve the efficiency of information interaction under the premise of ensuring the connectivity of the entire UAV communication network.Using potential energy function and neural network fast terminal sliding mode control technology,the adaptive neural network terminal sliding mode controller is designed for position outer loop and attitude inner loop respectively,which avoids the collision problem during formation flying and formation transformation,and realizes fast and stable tracking of formation trajectory.By choosing the controller parameters reasonably,the upper bounds of the position and velocity errors of the multi-quadrotor UAV formation system are obtained on the premise of ensuring the stability of the system.Finally,the simulation results demonstrate the effectiveness of the proposed control scheme.Finally,an adaptive sliding mode time-varying formation control problem based on state observer is studied for multi-quadrotor UAV systems under directed switched communication networks.By designing a distributed time-varying formation control protocol,the UAV can communicate only with the neighbor position information,which can get rid of the dependence on global information.Due to the existence of underactuation characteristics,the position auxiliary controller is designed to solve the attitude target information and control thrust.At the same time,a neural network observer is designed to observe the unmeasured information of the system,and the observation value is fed back to the adaptive sliding mode controller in real time,which improves the robustness of the UAV system.Then the formation system of quadrotor UAV is analyzed by Lyapunov theorem,and the error of multi-UAV formation is proved to be bounded and converges to the zero point.Simulation results show that the proposed control method can achieve time-varying formation control for multi-UAV systems,and the validity of the theoretical results is verified.