Research on Integrated Navigation Algorithm Based on Radial Basis Function Neural Network

The Qinghai-Tibet Railway is located on the Qinghai-Tibet Plateau in China. Due to its harsh climatic conditions and geographic environment, the line currently uses the Global Position System (GPS) to achieve train positioning. However, the navigation satellites are easily blocked by obstructions such as tunnels, resulting in a decrease in positioning accuracy. The conventional solution is to use multiple sensors such as an inertial navigation system (INS) for data fusion. But in the INS/GPS integrated navigation system, when the GPS signal is lost, the navigation accuracy will still decrease. In order to solve this problem, this paper proposes a radial basis function (RBF) neural network-assisted integrated navigation filtering algorithm. When the two systems are working normally, the measured raw information is used to train the RBF neural network. When the GPS fails, the measured and calculated values of the inertial navigation system are input to the trained network model to obtain the predicted error value, which is corrected to the inertial navigation system to obtain the final navigation information. The MATLAB simulation results show that when the short-term GPS information is missing, the integrated navigation system assisted by the RBF neural network has better estimation accuracy.