Research And Implementation Of Integrated Navigation Technology For Satellite Signal Transient Outage

The GPS/SINS integrated navigation technology has become the preferred choice for the majority of mobile devices and applications because of its ability to provide high accuracy navigation and positioning information. But the GPS satellite signal is very weak, vulnerable to external environment, especially when the GPS receiver is in the tunnel, signal blocked area, dense high building or dense forest, the GPS signal is likely blocked,resulting in GPS signal's outage. At this time, the navigation system's output accuracy will decline rapidly. In order to solve the problem, this paper firstly designs low-cost GPS/SINS integrated navigation hardware platform, analyzes the navigation information's feature when the GPS signal is available. When the GPS signal is unavailable, the neural network prediction model is designed to predict and compensate the output error of strapdown inertial navigation. This article has done the following works:Firstly, this paper researches the principle and the composition of low-cost GPS/SINS integrated navigation. The SINS navigation process is mainly analyzed. The integrated navigation hardware platform is designed by using MEMS inertial sensor-MPU9250 and Ublox GPS module. The data processing program is designed to eliminate the error of the sensors. The strapdown inertial navigation algorithm and the Kalman data fusion algorithm are respectively designed;Then, the application method of neural network model is studied in integrated navigation. Aiming at the problem of GPS signal outage, a static BP neural network and a dynamic Elman neural network model are designed to learn the SINS error features.When the GPS signal is missing, the trained neural network is used to predict the output error of SINS and compensate the output of SINS just like the GPS;Finally, the designed navigation hardware module is used to measure on a certain road, and the collected navigation data such as position, velocity and attitude angle are preprocessed to form the training samples for the designed neural network. For the situation that the GPS signal is missing for 30s and 50s, the trained-well network is used to predict the navigation data. The results show that the position output error based on the neural network prediction is much smaller than that based on an alone SINS, and the prediction result of Elman neural network is better than that of BP network.