| The small quadrotor UAV is a flexible,portable,and low-cost flight device that is currently used in aerial photography and cargo applications.However,all current applications require artificial manipulation of the drone’s flight action,which requires additional training for "flying hands." In the usual follow-up process,since the "flying hand" not only controls the flight path of the drone,but also aligns the carried pan-tilt camera with the target,the operation often depends on the person’s own perception of the environment.It is difficult to achieve the desired result.Therefore,how to control the intelligent flight technology of drones becomes a key issue to be solved.In this paper,a tracking control system for four-rotor UAV is designed.Firstly,the relative position of the ground target and the drone is solved by the perceptual algorithm.Then the hierarchical control algorithm is designed to realize the tracking control of the whole system.Finally,by constructing the actual flight platform,the algorithm is integrated into the processor to realize the tracking flight of the quadrotor UAV to the ground moving target.The main research contents are as follows:(1)In view of the complex variability of the flight environment of the four-rotor aircraft,starting from the force analysis of the propeller,combined with the relevant knowledge of the attitude description,and selecting the appropriate Euler angle sequence rules,the propeller provided in the engineering application is saturated.The characteristics are taken into account to establish the Newton Euler mathematical model as a transfer function of the control system,making flight control more suitable for different flight environments.(2)A set of visual perception algorithms is designed to obtain the relative position of the quadrotor UAV to the ground moving target.Firstly,the tracked target is identified by color recognition and shape recognition.Then,according to the correlation,the tracked target is visually tracked.Finally,by establishing a pinhole imaging model and establishing an auxiliary surface,the relative of the target and the drone is calculated.position.(3)According to the established dynamic equation,a hierarchical controller is designed to realize the tracking process control of the drone.The system is linearized by feedforward compensation,and the attitude time-varying signal coupled in the height subsystem is predicted by Taylor expansion,which reduces the hysteresis caused by directly sampling the attitude coupling signal.Using the method of saturation weight assignment,the invalid increment problem caused by the superposition saturation in the attitude control signal and the height control signal is converted into a priority problem,and the system disturbance due to the invalid increment is eliminated.At the same time,for the tracking problem of moving targets,the time-optimized tracking controller is designed with the position equation and velocity equation of the tracked target as the final state constraints.(4)Using the DJI flight platform of DJI,the flight system was designed,and the sensing algorithm and control algorithm were coded and integrated into the control system.After the simulation was verified by modeling,the visual perception experiment and cloud of the object were carried out.The tracking experiment and flight tracking experiment verify the effectiveness of the tracking control system designed in this paper. |
