Data-driven Control For Trajectory Tracking Of Fixed-wing UAV

Due to the high speed,strong maneuverability and long endurance,fixed wing Unmanned Aerial Vehicle(UAV)has attracted lots of attentions in recent years.However,it is very difficult to construct one accurate mathematical model for fix wing UAV,because of some nonlinear properties,such as under driving and strong coupling,etc,then some difficulties may appear in controlling fixed wing UAV for model based control strategy.Moreover,the susceptibility to external disturbance and the complexity of the flight environment are also two important factors in controlling process for fixed wing UAV.More specifically,trajectory tracking control is the basis for fixed wing UAV to complete the mission of target tracking and cruise,as the flight safety and mission completion will be greatly influenced by the trajectory tracking effect.Generally,it is well known that the essence of controlling fixed wing UAV to track the desired trajectory is to achieve the speed control and attitude angle control.Here the main contributions of this paper are formulated as follows.(1)Firstly,one nonlinear mathematical model is constructed for fixed wing UAV through the first principle method,then the principle of small perturbation is proposed to expand this nonlinear mode.After neglecting the high order terms,one approximated linear model is yielded for fixed wing UAV.(2)Secondly,to avoid errors,coming from the inaccurate model,data direct control method is applied to design the classical Proportion-Integral-Derivative(PID)controllers in closed loop and dual closed loop system,which are benefit for speed control and attitude angle control respectively.The controller parameter vector in PID controller is identified through the recursive least squares algorithm.Furthermore,Lyapunov stability theory is introduced to establish one parameter adaptive adjustment mechanism,being able to suppress the external disturbance.(3)Thirdly,the dual closed loop system or cascade system is considered to coincide with attitude angle control.Consider two PID controllers,existing in inner loop and outer loop for attitude angle respectively,data driven control strategy is also used to design these two unknown PID controllers by solving two minimum optimization problems without any knowledge about the considered system.During the whole optimization process,one simple first order gradient algorithm is used to solve the optimization problem,and its convergence is analyzed too.(4)Finally,lots of simulation examples are given to prove the efficiency of our considered data driven control strategy in tracking the desired trajectory.