Finite-time Attitude Control Of Spacecraft Based On Adaptive Robustness And Neural Network Approximation

The development of modern aerospace industry requires spacecraft to have high attitude control accuracy,and meanwhile it also requires the attitude of the spacecraft to converge to the equilibrium point as soon as possible.However,on the one hand,when the spacecraft works in orbit,it will suffer from various disturbances from inside and outside,including the disturbance torque from the outside as well as the influence caused by the uncertainties of the spacecraft model parameters,thus affecting the control accuracy of the spacecraft;on the other hand,many current researches can merely guarantee asymptotic stability at the equilibrium point of the system,which means that the attitude error requires an infinite time to converge to the equilibrium point theoretically.In view of the problems above,the finite-time attitude control problem of spacecraft with internal and external disturbances will be deeply studied in this paper.The main contents include the following aspects:Based on the traditional sliding surface,a fractional power term is i ntroduced into the sliding surface,and a novel terminal sliding surface suitable for spacecraft attitude tracking is designed in this paper.In order to solve the singularity problem that may exist in the novel terminal sliding surface,it is improved and a non-singular terminal sliding surface is obtained.Theoretical analysis shows that there is no singularity problem in the modified terminal sliding surface,and it can achieve finitetime convergence to the equilibrium point.A finite-time spacecraft attitude controller based on adaptive robustness is designed in this paper by using the modified non-singular terminal sliding surface.By introducing three groups of adaptive parameters,real-time online estimation of the total disturbance upper bound is realized,and the estimated total disturbance upper bound is applied to the robust term in the controller,so that no priori disturbance information is required when the controller is designed.Theoretical analysis proves the finite-time stability of the controller,and numerical simulations show that the controller has strong robustness and realtively high control precision.A finite-time spacecraft attitude controller based on neural network approximation is designed in this paper by using the modified non-singular terminal sliding surface.Through the Chebyshev neural network,the approximation of the total disturbance is realized,and the total disturbance is compensated in the controller,which thereby greatly avoids the influence of internal and external disturbance on the attitude control.Theoretical analysis proves the finite-time stability of the controller,and numerical simulations show that the controller achieves high-precision attitude control.In this paper,according to the control performance of the finite-time spacecraft attitude controller based on adaptive robustness and the one based on neural network approximation,the switching function is introduced to make the two controllers play a dominant role in different control stages respectively.Thus,a finite-time spacecraft attitude controller combining adaptive robustness and neural network approximation is obtained.Theoretical analysis proves the finite-time stability of the controller,and numerical simulations show that the controller has fully inherited the advantages of the two controllers above and it has both fast convergence speed and high control precision.