Research On The Embedded Platform For Combined Navigation Of Quadrotor Aircraft

Integrated navigation technology has become a focusing topic with the development of four-rotor aircraft technology.The traditional integrated navigation platform has disadvantages of low accuracy,weak fault tolerance and poor real-time performance.In order to meet the requirements of high-precision navigation of aircraft in complex environments,it is necessary to design an embedded integrated navigation platform which is capable of achieving high-precision navigation under complex and changeable environmental conditions.This thesis aims at designing an embedded platform for integrated navigation of four-rotor aircraft to realize data acquisition and multi-sensor navigation data fusion in a complex environment,and solve the problem of error accumulation of inertial navigation module when GPS(BeiDou)is invalid which can improve the accuracy and fault tolerance of the navigation platform.Firstly,the overall framework of embedded hardware platform and Linux software based on S5PV210 processor and multi-sensor integrated navigation is designed.The system software modules are designed specifically to complete the design and transplantation of uboot,kernel and rootfs.The multi-threaded design of the software platform and the preparation of the underlying driver of the data acquisition system are also completed.The algorithm is based on the federated filtering model,using the distributed Kalman filter method for INS/GPS/monocular vision data fusion and using the indirect Kalman filter method to implement INS/GPS and INS/monocular vision based on the error model as two sub-filters.The main filter implements global optimal estimation according to the method of weighting the optimal coefficients.Finally,the airborne test and comparison experiment of the embedded integrated navigation platform were carried out.It further verifies that the monocular vision sensor has a good correction ability for the error accumulation problem of the inertial navigation module under the condition of GPS invalidation.It also verifies that the platform accuracy and fault tolerance performance of multi-sensor embedded integrated navigation based on monocular vision is significantly better than single or arbitrary two-integrated navigation system.This research has practical reference significance for the research of integrated navigation related technology.