An adaptive fuzzy logic quaternion scaled unscented Kalman filtering for inertial navigation system, GPS and magnetometer sensors integration

In this paper, we present a technique based on fuzzy logic to improve the performance of the inertial navigation system integrated with GPS, and magnetometer. The proposed fuzzy technique is primarily used to predict position and velocity measurements during GPS outage signals. As long as the GPS measurements are available, the Q-SUKF of INS/GPS/MAG (MAG: magnetometer) integrated system operates efficiently and provides precise navigation states estimation. Nevertheless, during GPS outage signals, the proposed fuzzy technique is adapted to the Q-SUKF to obtain the (A) (FL) QSUKF (Adaptive Fuzzy Logic Quaternion Scaled Unscented Kalman Filter) in order to correct the degradation of the performance of the algorithm. The adaptive fuzzy logic attributes values to the measurements covariance matrix in order to determine the gain of the filter. It will decrease the measurement noise variance of the Kalman filter and then improves eventually the accuracy of the integrated navigation system states estimation. Finally, an experimental part on the use of the proposed fuzzy technical with the Q-SUKF has been validated. Several GPS outages with duration of 30s have been simulated to study the behavior of the proposed filter. In addition, an initial attitude error of 60 degrees is given in each axis to test the robustness of the filter proposed under large attitude errors. The results of the experimental validation have shown the effectiveness and the significant impact of the (A) (FL) Q-SUKF in the reduction of the drift errors estimation of the position and velocity in case of GPS outages in the tested scenarios.