The Research On Intelligent Control Technology Of Four-rotor Aircraft

The quad-rotor aircraft is quadrotor aerocraft is a kind of multi-axis aerocraft with low drive,strong coupling and highly nonlinear system,rotating through four rotors to maintain a stable posture and the movement of the aircraft.Four rotor aircraft because of its simple structure,low maintenance cost,more and more welcomed by the market,but how to solve the underactuated nonlinear and strong coupling features,has been the research key points and difficulties of attitude control,nowadays the research emphasis of spacecraft attitude control algorithm is neural network,main footwork,such as fuzzy control,these studies have made great progress,but these algorithms for vehicle model has great dependence,facing the situation in the practical engineering will be more complex,because the aircraft in flight will be affected by the outside world wind,aerodynamic,gyroscopic effect,air and so on a series of interference caused by vibration,and it is very difficult these interference by conventional methods for modeling,on the other hand,the parameters of the aircraft itself also has a lot of uncertainty,such as flight load quality change,after another to calculate the rotational inertia of the vehicle is very difficult to even out the parameters and the actual model also there is a big gap,it is very difficult to build precise mathematical model,it is also a lot of problems existing in the study,although the simulation results is very good,but if applied to the actual system will find algorithm thoroughly failure,the root cause lies in the huge difference between the simulation model and actual model.Learned the modern control theory,this paper,intelligent control,such as curriculum knowledge,small four rotor aircraft with uncertainty as the research object,mainly from the following aspects has carried on the research of spacecraft attitude control:1.The aircraft model,first analysis of aircraft on the ground and the body coordinate system transformation relationship,and then by Newton's second law and moment of momentum theorem,from two aspects of translation and rotation stress analysis was carried out on the aircraft,the translational and rotational dynamics equation respectively,and thus get the aircraft model from two aspects of translation and rotation.2.PID control for aircraft respectively,RBF neural network and PID combination of RBF neural network adaptive PID control and sliding mode control,RBF networks and adaptive algorithm of RBF network adaptive sliding mode control algorithm is designed.PID control does not need to know the model information.The control task can be accomplished by adjusting PID parameters,but the stability is poor.The adaptive PID control of RBF neuralnetwork realizes the tuning of PID parameters by applying the proposed Jacobian algorithm value to the PID parameter.RBF neural network adaptive sliding mode control is based on the unknown function of on-line approximation system model,and it is designed to estimate neural network weight and model unknown parameters.The last of this algorithm by MATLAB/SIMULINK simulation,the results show that RBF neural network adaptive sliding mode control is better than RBF neural network adaptive PID control and PID control can achieve better trajectory tracking,and adjust the volume of time shorter,overshoot smaller and stronger anti-interference ability,in the case of aircraft quality change at the same time,the robustness of the controller.3.The overall design of the quadrotor aircraft prototype was completed from two aspects of hardware and software respectively,and outdoor flight test was conducted to test the attitude stability of the aircraft.The aircraft finally designed successfully completed the take-off,landing,suspension and return missions.The attitude information stored in Pixhawk indicates that the aircraft can track the expected attitude Angle.