Research On Intelligent Attitude Control Method Of UAV Based On Three-degree-of-freedom Helicopter Model

Rotor unmanned aerial vehicles have been widely used in military and civil fields,as they have simple structures,superior maneuver abilities and they can take off and land vertically.The 3-DOF helicopter experimental platform can simulate the flight attitude of a two-rotor helicopter.It is an auxiliary tool for studying basic control theories and flight control methods of helicopters.The 3-DOF helicopter system belongs to a kind of high-order,strong coupling and nonlinear system,so effective control methods must be taken to get good control results.This paper studies the intelligent attitude control methods for the 3-DOF helicopter system,and the main work is as follows:(1)According to the characteristics and features of the 3-DOF helicopter system,the mathematical model of the 3-DOF helicopter is established based on the dynamic equations of three axes.Then the PID controllers are designed for attitude control of this system.Experimental results show that although the PID control method can meet the control demands,the overshoot of system is very high and the adjusting time is very long.The control consequence is not ideal.(2)Aim to overcome the shortcomings of the PID control method,the fuzzy PID controller is designed.This controller determines fuzzy rules based on expert experience and experimental datas,and then uses these rules to adjust the PID parameters in order to meet requirements.Experimental results show that with the fuzzy PID control method,the helicopter has smaller overshoot and shorter adjusting time.At the same time,the anti-interference ability of the system is increased.(3)Due to the existence of nonlinear and strong coupling problems in system,the PID Neural Network(PIDNN)controller is designed.The controller determines initial values of network connection weights according to the basic principles of the PID control method.Then,error back propagation algorithm is used to adjust the connection weight values.The mean square deviation between actual outputs and ideal outputs is taken as the optimization objective function.The goal is to achieve the optimal objective function value by training and learning constantly.Experimental results show that with the controlling of PIDNN,the helicopter has good dynamic characteristics and good abilities of decoupling.(4)This paper designs the PID controller,the fuzzy PID controller and the PIDNN controller of real-time controlling experiments based on the Googol 3-DOF helicopter experiment platform.By analyzing real-time performances under different control methods,advantages and disadvantages of these controllers are discussed and advice for practical engineering is given.