基于BP神经网络辅助的车载组合导航算法研究

针对INS/GPS组合导航系统在GPS信号被遮挡时，GPS接收机失锁导致导航精度迅速下降的问题，提出了基于BP神经网络辅助的组合导航算法。即在GPS信号锁定的时候，采用卡尔曼滤波对INS/GPS信号进行数据融合得到实时的精确位置，同时利用组合导航输出信息对BP神经网络进行实时在线训练；一旦GPS失锁，利用之前训练好的神经网络对INS系统进行误差补偿，解决精度迅速下降问题。通过跑车实验证明，速度精度在0.2m/s以内，位置精度为25m以内，该算法对INS/GPS组合导航系统有效。

Research on Integrated Navigation Algorithm Based on BP neural network

Aiming at the problem that the GPS receiver is blocked when the GPS / GPS integrated navigation system is blocked, the navigation accuracy of the GPS receiver is reduced rapidly, and an integrated navigation algorithm based on BP neural network is proposed. That is, when the GPS signal is locked, the data fusion of INS / GPS signal is realized by Kalman filter, and the BP neural network is trained in real time by using the integrated navigation output information. Once the GPS is lost, Neural network on the INS system error compensation to solve the problem of rapid decline in precision. Experiments show that the speed accuracy is within 0.2m / s and the position accuracy is less than 25m. The algorithm is effective for INS / GPS integrated navigation system.