Research On Quadrotor Attitude Control Based On Multi-sensor Fusion

Because of quadrotor's flexible use,vertical take-off and landing,and its low cost of using,it's gradually applied to various fields such as precision agriculture,smart city and military reconnaissance.More and more scientists contribute to research in this field.During in a complex application environment,precise attitude and position data must be obtained to achieve precise control of the quadrotor.However,the sensor signal of the quadrotor may be disturbed by external factors such as temperature or wind field.It is difficult to obtain accurate position and attitude information,which makes it difficult to achieve precise attitude and position control.In addition,for low-cost quadrotor,inertial navigation has low accuracy and large cumulative error,and inertial navigation cannot provide accurate position information.When the satellite signal is good,the position information of the quadrotor can be calculated through the analysis of multiple satellite signals.However,in some situation,the satellite information is not sufficient to get quadrotor's position information,which cannot achieve its own navigation.In this paper,the research is to improve the perception and adaptability of the quadrotor to the environment.The main work is as follows:Firstly,the mathematic model of the quadrotor is established,and the sensor calculation principles of accelerometer,gyroscope,magnetometer and so on,which are commonly used for attitude measurement,are analyzed.According to the characteristics of various sensors,complementary filtering algorithm and extended Kalman filter algorithm are used to fuse the multi-sensor measurement data to obtain more accurate attitude and altitude information,so as to realize the attitude estimation of the quadrotor.The simulation results show that the two fusion algorithms are effective,and the latter has better performance.Secondly,due to the high nonlinearity and time-varying of the position and attitude control system of the small quadrotor,the conventional PID controller has poor control effect and is difficult to adapt to complex flight environment.In order to overcome this problem,the single neuron PID control method is used to improve the control performance of the quadrotor by adjusting the weight of single neuron in real time.In order to avoid the problem of large sample size and large optimization parameters,a fuzzy single neuron PID control strategy is proposed to improve the dynamic performance of the quadrotor control by automatically adjusting the gain coefficient.Based on MATLAB/SIMULINK,a simulation system of the quadrotor is built.The simulation results show that the attitude control of the quadrotor based on the proposed method has the advantages of smaller overshoot and faster response speed.Furthermore,in order to reduce the dependence of the quadrotor on the environment,when it is unable to receive the satellite positioning signal,the position information of the quadrotor is acquired by using the visual sensor on board.The information is used to analyze the feature points of the ground image,and the visual recognition information is introduced into attitude and position control of the quadrotor to realize its attitude and navigation control in some environment.Finally,the indoor test environment is built for the small quadrotor.Based on PIXHAWK,the hardware platform of the flight controller is constructed,and the program PX4 of the open source quadrotor is optimized.The simulation and experimental results verify the correctness of the theoretical analysis and the effectiveness of the proposed control strategy.