Research on Improving the Elman Neural Network for the INS/GNSS Integrated Velocity and Positioning Algorithm

Speed and positioning technology is a crucial foundation and key technology for the operation control and safety assurance of maglev trains. In this study, an improved bee algorithm (IABC) optimized Elman neural network-assisted INS/GNSS combination positioning method is proposed to address the issues of satellite signal loss and INS/GNSS error accumulation, which can affect the accuracy of the system. When the satellite signals are reliable, the system adopts the INS/GNSS combination speed and positioning mode. The IABC is used to optimize the weights of the Elman neural network, thereby improving its prediction accuracy. The optimized Elman neural network is trained accordingly. When satellite signals are lost, the trained neural network is utilized to predict the measurement input of the Kalman filter, providing error correction values for the output of the inertial navigation system to enhance speed and positioning accuracy. Simulation experiments have demonstrated that the IABC optimized Elman neural network-assisted combination navigation system significantly improves the speed and positioning accuracy when satellite signals are lost.