Liquid tank modelling and attitude controller design of plant protection UAV

To improve the controllability of plant protection ummanned aerial vehicle (UAV) with eight rotors, the level of liquid in the tank is analysed and modelled during pesticide spraying process. As a result, formulas for the liquid mass and moment of its inertia are presented in terms of time, causing an accurate dynamical model in time domain. To validate the correctness and effectiveness of the model, attitude controllers are designed by using adaptive backstepping terminal sliding mode control technology. Firstly, the typical backstepping control method is used, then the terminal sliding mode control is introduced for the last state to converge the ideal value in a finite period of time, finally, an adaptive control law is established to eliminate uncertain disturbances. All the designed controllers have met the Lyapunov stability criterion. These proposed controllers are simulated and tested, whose results are compared with those of fuzzy PID attitude controller. The results show that the proposed dynamic model make the adaptive backstepping terminal sliding mode controller applicable for the plant protection UAV. Compared with Fuzzy PID controllers, the effectiveness of attitude controller is improved greatly, the errors in attitude angle are reduced by 25.57%, 24.21% and 19.41% in the experiment, respectively. Additionally, the controllers are insensitive to random disturbances and subject to a nicer robustness property.