Research On UAV Integrated Navigation System Based On Multi-sensor Information Fusion

The UAV integrated navigation system based on multi-sensor information fusion can real-time detect the UAV’s position,speed,attitude and other flight status information in a variety of environments.Compared with a single navigation system,it has higher environmental adaptability,and is the foundation and core technology for the UAV to achieve autonomous flight and complete related tasks.At present,the navigation system used by UAV is usually a single navigation mode with multi-sensor configuration.When UAV is in a complex environment,it is usually difficult to obtain accurate and reliable navigation information.In order to improve the reliability of UAV navigation system in complex environment,a new integrated navigation system is designed in this paper,which combines monocular camera,GPS,barometer and ins.The main work of this paper includes:(1)Aiming at the outdoor open environment where GPS can be used,an integrated navigation system with STM32F407VGT6 processor is designed,which is equipped with accelerometer,gyroscope,geomagnetic,barometer and GPS sensors.In this system,with time stamp alignment and fault diagnosis technology as the core,in view of the delay of GPS relative to INS data,firstly,a predictor is designed to compensate the delay of GPS data,so that INS data and GPS data can be synchronized in time;secondly,a fault diagnosis judge is designed to measure data(GPS and barometer data)and predict data(acceleration count)According to)carry out fault diagnosis to prevent the system from working normally due to the short disappearance or jump of measurement data.Finally,the time stamp alignment technology and fault diagnosis technology are integrated into the improved extended Kalman filter to estimate the position,velocity and attitude information of UAV.(2)In the environment where GPS can not be used,the integrated navigation system with STM32F407VGT6 and industrial computer dual processor is designed.The system is equipped with monocular camera,gyroscope,geomagnetic and accelerometer sensors.In this system,based on slam technology,the front-end uses multi-layer optical flow to track feature points,and the back-end uses nonlinear optimization method to fuse camera data and IMU data,so as to obtain the location information of UAV.The attitude information is calculated from accelerometer data,gyroscope data and geomagnetic data by extended Kalman filter algorithm.Because the camera and IMU can only measure the relative position information of UAV,when the UAV runs in the GPS signal area for a long time,there will be accumulated error in the calculated position information.In order to reduce the accumulated error of position estimation,the loopback detection function is added.