Research On Control Algorithm Of Quad-rotor Aircraft Based On Neuro-fuzzy PID Control

The four-rotor aircraft is not easy to be hindered by the space constraints of the site to take off and landing,and it is flexible and highly controllable.With the progress and development of science and technology,the organic combination of intelligent equipment and quadrotor makes its own value fully utilized and has a wide range of application prospects.UAV traces will be found in both the civilian and military fields.Quad-rotor aircraft performing flight tasks often have problems such as low stability attitude accuracy and poor resistance to interference.Therefore,it is necessary to carry out extensive and meaningful research on the control algorithm of the quadrotor.The main content of the paper is as follows:First,this article analyzes the basic system structure and working principles of the quad-rotor,and establishes the flight dynamics system model of the quad-rotor using Newton’s Euler principle.The complex mathematical model of the quadrotor is used for linear optimization of the dynamic model by the small disturbance method,and the space state equation is established.After decoupling,the transfer functions of the four channels of yaw,pitch,roll and vertical can be obtained,which will provide the controlled object for the subsequent chapters.Secondly,the control parameters of the quadrotor under the classic PID control algorithm are preliminarily set,and the fuzzy PID controller is designed on the basis of the PID controller.The fuzzy rules,membership functions,quantization factors and other parameters of the quadrotor were tested for many times,and the simulations were carried out in Matlab/Simulink environment to compare the two control algorithms.Finally,it integrates the fuzzy rules set by the fuzzy PID controller and the existing problems in the membership function domain,and introduces the neural network model to adjust the membership function and fuzzy rule parameters of the fuzzy PID controller by using its self-learning ability.Construct training and test data sets based on the data samples of the fuzzy PID control system.The training algorithm is composed of a combination of BP algorithm and least square method,and the trained adaptive neuro-fuzzy control system(ANFIS)is combined with PID.Into the mathematical model of the quadrotor,and use different input variables to simulate the flight instructions encountered by the quadrotor during the flight.In order to fully verify the effectiveness of the neuro-fuzzy PID controller,an anti-interference test of three algorithms is carried out.