Control Strategy And Application For Vision-based Autonomous Landing Of Micro Aircraft Vehicles

Quadrotor is a kind of novel appearance, performance excellence UH for unique ability to fly,have been widely used in the aerial photography. Autonomous landing is one of the importantfunctions of UAV flight control system, which influence the UH autonomous ability and the level ofthe engineering application directly. X450quadrotor as the platform, this paper mainly studies itsflight control and the autonomous landing of the control strategy based on vision navigation.Firstly, the established quadrotor model is trimmed and linearized, the result show that themathematical model is not accurate. According to the tasks the quadrotor need to accomplish,quadrotor will adopt GPS navigation in cruise and point-to-point flights, only the vision navigationwill be adopted in the autonomous landing task.In order to solve imprecision problem of model, avoid air turbulence, active disturbancesrejection control (ADRC) method is used in inner-loop, and the classical PID control approach isapplied in outer-loop. Simulation results show that the control plan is effective, and reducing themodeling workload, and the quadrotor is able to reach destination, achieving attitude and positioncontrol. According to the difference between the visual and GPS navigation system, visual landingusing separate control algorithm is putted forward.According to the delayed signals and mismatch between vision and control sample frequency,this paper studied the visual information predication and interpolation. The relative position signalsbetween quadrotor and land field to construct AR prediction model. prediction method of aircraft andlanding zone relative position for processing AR model is established, then using forgetting factorrecursive least squares method to estimate the parameters of AR model, solve the problem of delay,the simulation results show that the prediction method has a good effect.Finally, this article give the quadrotor simulation of whole control system. The simulation usingthe Stateflow toolbox for flight task management, but also uses flightgear for visualization ofsimulation process. Simulation results show that the designed control system has good effect.