Flocking Control Of Rotor Aircraft Based On Neural Network

Flocking is a kind of group movement behavior, which can be described as follows:after group from the state of separation to together, they move together in a specific direction,and maintain a particular Rules at same time. As for some tasks, group flocking has unparalleled advantage with a single motion. this has the very important guiding significance and application to coordinated control issues. Especially,when it is applied to having a actual physical model, the study of the flocking movement has a very important practical significance.On the other hand, with the rapid development of new materials and control technology in recent years, the rotor UAV has become a research hotspot in the field of UAV research,due to its own advantages, and a certain accumulation in the control theory has been done. However, for the control of the full model, as well as multi machine control is still the weaknesses of the study. Therefore, the main objective of this paper is based on the study object of quadrotor unmanned aerial vehicle(UAV), by researching flocking controller with quadrotor UAV model, achieve flocking behavior of quadrotor UAV Groups.In this paper, the problem of quadrotor unmanned aerial vehicles(UAVs) flocking control is divided into three problems. The first one is the control of quadrotor UAV, the second one is flocking control based on particle model. And then, Combined with dynamic characteristics of the rotor UAV, complete the flocking control of quadrotor UAVs.For the highly nonlinear, strong coupling and underactuated, an effective controller for quadrotor UAV is not easy to obtain, especially in conditions of the parameters variations and external disturbance. In this paper, an adaptive inverse model control system(AIMCS) is proposed for this plant. Two BP networks are used in control system to achieve the model identification and control. Both offline trained and online learned are used to ensure the fast learning and the robustness. The convergence of the learning algorithm is proved based on Lyapunov function. At last, a quadrotor UAV control simulation based on the full model shows the superiority and robustness of the control system.After completing the control system of the rotor UAV, the flocking algorithm is the key point of rotor UAV flocking control. In this paper, using the energy function translates the three rules of Reynolds into a mathematical problem: through establish appropriate artificial potential function, the particle aggregation, collision avoidance and speed consistency problems are transformed into potential energy function minimization problem,and then design the controller based on Liapunov method, so that the artificial potential energy function is minimization, so as to realize the particle model flocking control. In the design process, saturation input is considered,so that it is effective using the control system of the rotor UAV.using dynamic characteristics of the rotor UAV, Output of flocking control algorithm will be translated into the target of attitude of quadrotor UAV. Combined with the adaptive inverse control system proposed in this paper, the flocking control of quadrotor UAV is completed.