Translated Abstract
In the current information age, any single navigation method is difficult to satisfy human＇s requirements for navigation accuracy. Each navigation method has its own advantages and disadvantages, so integrated navigation has become one of the most promising technologies in the navigation field, and it has always been one of the hotspots of scientific research. The strapdown inertial navigation technology has the advantages of good concealment, fast updating frequency and good stability, but the inertial navigation system can not work for a long time due to the drift of inertial sensors. The GPS has the advantages of full time and all weather, and can easily locate, measure time and measure speed in the world. However, the GPS data update frequency is slow and easy to be disturbed by external information. . Considering the advantages and disadvantages of the above two navigation methods, GPS/SINS integrated navigation has been widely applied. This paper mainly focus on the integrated navigation system of GPS/SINS.The main takes are as following:
This paper studies the principle of the strapdown inertial navigation and the principle of calculation, and establishes its attitude, velocity and position error equation in the navigation coordinate system.
The mathematical model of the trajectory of the motion state of the carrier is generated. And thus the actual output data of the gyroscope and accelerometer is produced. The Kalman filter mathematical model of the integrated navigation system under the loose combination method was established, and simulation experiments were conducted on the positioning error of the pure strapdown inertial navigation system and the integrated navigation system. The results showed that the integrated navigation positioning effect was better than the strapdown inertia navigation system.
For the integrated navigation system of GPS/SINS, the error of the positioning precision of the single inertial navigation system is large when the GPS signal is unlocked, and a radial basis function neural network is proposed to assist the Kalman filter. In this method, the angular velocity and acceleration of the carrier are used as input, the measurement is taken as the output, and the RBF neural network is trained before the GPS failure. The network is used to predict the view measurements during the failure of the GPS and then the Kalman filter is then carried out to keep the integrated navigation system in a relatively stable state. The simulation results show that the precision of navigation information is higher, the system is more stable, and the training time is shorter than the BP neural network which is more widely used. It can meet the requirement of real-time positioning of integrated navigation.
Translated Keyword
[Integrated navigation, Kalman filter, Loose combination, Neural networks]
Corresponding authors email
