Design Of Fractional Kalman Filter Algorithm

In recent years,UAV has been widely used in civilian applications.Especially the vertical take-off and landing UAV can be widely used in geological mapping,geological exploration,forest protection and transportation of goods.In this paper,a fractional Kalman filter,which is based on the civilian vertical take-off and landing UAV,is designed for flight attitude measurement system.This paper discusses the background and significance of the research,and it briefly introduces the research status of the aircraft attitude calculation algorithm,the development process of the fractional Kalman filter,and the theoretical basis of inertial navigation system and attitude calculation.The theoretical basis,which including the characteristics and classification of inertial navigation system,the definition of the common coordinate system,the definition of the attitude angle,three representation methods of attitude information and the relationship between the three methods,and the characteristics of three kinds of attitude calculation methods,etc.This paper briefly analyzes the attitude measurement system of UAV,and specifically introduces the working principle of gyroscopes,accelerometers,and magnetometers.And according to the previous introduction,this paper proposes the STM32F103ZET6 microcontroller as the main control chip,the MPU6050 chip and the HMC5883 L chip as attitude measurement sensors,and designs the hardware circuit and system software framework,and give the design principle and circuit schematic diagrams of microprocessor,attitude sensor,communication protocol and power module.By understanding the traditional discrete Calman filter and the fractional calculus theory,the Grunwald-Letnikov differential definition is introduced into the integer order system.Based on the difference idea,the integer order system is generalized to the fractional order system,and combined the Calman filter and the fractional order system,the integer order Calman filter is extended to the fractional Calman filter.The simulation results show that the fractional Kalman filter has a good filtering effect.Fractional Kalman filter algorithm is designed based on the established system state equations and observation equations.The algorithm merges the attitude data collected by the three-axis gyroscopes,accelerometers,and magnetometers to obtain more accurate attitude information.According to the designed algorithm,the attitude calculation is verified on the hardware system,and the changes of the attitude angle are displayed by the third-party PC software.The experiment is divided into two parts:(1)dynamically detect attitude angles,then compare pre-filtered and filtered attitude angle waveforms;(2)statically preset the angle,and then compare the attitude angle value output by the attitude measurement system designed in this paper with the value output by other attitude measurement systems.Experimental results show that the system can output relatively stable attitude data for a long time.