Relay Balance Control Of Double Quadrotors Inverted Pendulum

With the development of science and technology,quadrotor aircraft is more and more widely used.It can be equipped with cameras,slings,manipulators,etc.it can be extended to all walks of life to engage in all kinds of dangerous work that cannot be done by human.It can be predicted that it will be needed to participate in more complex tasks in the near future.As a typical nonlinear underactuated system,it is difficult for quadrotor aircraft to carry out accurate mission control in complex environment,especially to control its interaction with external objects,such as changing battery items in the air.In order to meet the complicated control requirements,this paper research the mixed balance manipulate device of quadrotor and inverted pendulum,deeply analyzes the balance prerequisites and control methods,designs the simulation test platform,takes quadrotor as the driving platform to manipulate the inverted pendulum balance,and uses reinforcement getting to know algorithm to complete the relay simulation scan of inverted pendulum in the air.The foremost research work of this paper consists of the following aspects.(1)The mathematical models of four rotor and inverted pendulum are established respectively.According to their dynamic relationship,the system model of four rotor inverted pendulum is established and simplified.The control input relationship and control conditions are constructed.The conclusion that the inverted pendulum can be balanced by changing the attitude angle of four rotor is obtained,which reduces the control difficulty of complicated machine.It is handy for the subsequent evaluation of the experimental process.(2)The rotating coordinate system is established respectively,and the balance environment of four rotor inverted pendulum is designed in two cases.The balance conditions of fixed-point hovering and circular motion are obtained.The LQR controller is used to simulate on the designed simulation platform,and the stability control of four rotor inverted pendulum gadget is realized,which provides basic conditions for subsequent relay experiments.(3)In this paper,the relay process decomposition design of four rotor inverted pendulum system and the different motion states of the inverted pendulum light bar are analyzed from the perspective of energy conservation.The formulas for calculating the optimal connecting point and connecting time are derived.The velocity and attitude conditions that the four rotor can meet are analyzed.The throwing motion of the four rotor is designed according to the first principles and simplification principles Based on this,the four rotor extension method is designed,and the variable acceleration method is used to control the four rotor maneuver,and the predicted path of the four rotor extension is obtained.The feasibility and reliability of the predicted path are verified by simulation.(4)This paper determines the curve description of the tracking path of the connecting rod,introduces the strategy gradient reinforcement learning algorithm to design the control learning strategy and network parameters to obtain multiple groups of training models,uses the gradient reinforcement learning algorithm based on the depth strategy to train the quadrotor eight character control,and obtains a good path following effect through data analysis.The trained agent is applied to the connecting rod experiment,and the simulation results are verified The path following effect is verified,and the relay experiment of inverted pendulum light bar is completed,which shows that the designed control algorithm verification platform meets the requirements of complex control of quadrotor inverted pendulum,and verifies the feasibility of the whole system design,which has a certain positive significance for the future control simulation field.