Data Fusion Algorithm for INS/GPS/Odometer Integrated Navigation System

INS/GPS/odometer are commonly integrated using a federated Kalman filter to provide a robust navigation solution, overcoming their weaknesses. However, the accuracy of federated Kalman filter is degraded in the condition that the statistical characteristics of noise don´t be known accurately. The method of federated Kalman filter is improved to perform the INS/GPS/Odometer integrated navigation in this paper. This method uses fuzzy adaptive Kalman filter to detect changes of the measurement noise statistical characteristics and correct them gradually. Meanwhile, weighted coefficient is used to describe the degree of confidence of sub-filters. Simulations in INS/GPS/odometer integrated navigation system demonstrate that the weighted coefficients of sub-filters with low confidence are decreased adaptively, and the accuracy is improved compared with the federated kalman filter.