High dynamic GPS/INS integrated navigation fusion algorithm assisted by neural network

The phenomena of the disconvergency of Kalman filter occurs in case of lock-lose or absence in GPS (global positioning system) navigation based on carrier phases in the high dynamic environment. We propose a new method to enhance the performance of GPS/INS (intertial navigation system) integrated navigation system during the GPS signal loss. The method is based on using a ANN (artificial neural netwok) to intelligently aid the GPS/INS integrated navigation system in the absence of GPS signal. The ANN is trained in every filter period during the GPS signal on; When the GPS is off, The output of the previously trained ANN is as the input of the same Kalman filter.The proposed enhanced GPS/INS can be used in the high dynamic environment of a autonomous carrier. A simulation indicates that the method is valid.