Research On Algorithm Of Multi-source Assisted Inertial Navigation System

With the rapid development of sensor technology and information technology,the traditional single navigation system positioning method will limit the applicable scenarios,and the reliability and accuracy of navigation system will also be affected.At present,the joint positioning of multi-source data sources has become a hotspot in navigation technology.Among many navigation systems,Strap-down Inertial Navigation System(SINS)has the advantages of autonomy and concealment.With the continuous development of inertial device technology,micro-electro-Mechanics System(MEMS)based micro-inertial measurement unit(IMU)has emerged,which has the advantages of small size,low power consumption and controllable cost.The advantages expand the application field of inertial navigation.And inertial navigation system can be used in micro or small tactical carriers.According to the advantages of inertial navigation system,a multi-source data fusion navigation system software platform based on inertial navigation is built,and the algorithm of inertial navigation system is studied.In this paper,based on a national defense key project,the algorithm of multi-source aided inertial navigation system is studied.The main tasks are as follows:(1)Introduced the related theory knowledge of strapdown inertial navigation system,including attitude updating algorithm,speed and position calculation algorithm Because the attitude calculation using traditional quaternion in the case of intense angular motion will have non-convertible error,it is necessary to study the coning error compensation algorithm,and based on the principle of minimizing the error of the algorithm,to optimize the coefficients of traditional double-sample method.In the simulation of classical coning motion environment,it is proved that the precision of the improved double-sample algorithm is slightly better than that of the traditional double-sample method.(2)The error types of MEMS gyroscope are analyzed.According to the non-linearity of random noise and the characteristics of susceptibility to test environment,it is impossible to establish a fixed model.Therefore,time series algorithm suitable for engineering is used to build the model,and Kalman filter is used to compensate the random noise with statistical characteristics.Because the noise characteristics are unknown,an adaptive algorithm is added to solve the measurement noise matrix.The residual error is used to estimate the measurement noise matrix in real time to improve the filtering accuracy.Finally,a method for evaluating random noise,Allance Variance(AV),is introduced.(3)Firstly,the basic sources and compensation methods of space-time registration errors are analyzed;secondly,the federated filter structure used in this paper is briefly introduced,and the trajectory generator is used to generate the required trajectories.The accuracy and fault tolerance experiments are carried out for the three information allocation factors,so as to verify that the third information allocation factor has the best performance in fault tolerance and is tested in normal accuracy experiments.The accuracy of the other two information allocation factors is almost the same.(4)The hardware and software design of micro-inertial navigation/satellite navigation system is introduced.The hardware includes the overall structure of hardware and the characteristics of Soc chip.The software includes the division of software functions,the overall flow of software,the module of inertial navigation solution and the delay compensation of information fusion.Finally,the reliability of the whole system design and the accuracy of system are verified by ground static test and field test.