Parallel Multiple Methods with Adaptative Decision Making for Gravity-Aided Navigation

Gravity-aided navigation is an effective navigation method for underwater vehicles. However, the distribution of the gravity field may affect the measurement errors of gravity anomalies and the precision of gravity-aided navigation. In this paper, the upper and lower thresholds of the gravity field standard deviation are computed by the statistical properties of the local gravity field to classify each grid in the gravity map into different levels. A parallel multiple methods with adaptive decision making (PMMADM) for gravity-aided navigation is proposed which incorporates the gravity anomaly measurements, particle filtering, and maximum correlation method into the observation equation of the extended Kalman filter. The algorithm autonomously selects the observation variables in the filter by combining the gravity field standard deviation at the current position of the carrier with a decision tree. This approach can combine the characteristics of different gravity matching algorithms, reduce the impact of random noise in the measurements, and improve the positioning accuracy of gravity-aided navigation. Physical simulation experiments demonstrated that the proposed gravity matching algorithm achieves the high navigation accuracy and long-term stability in different gravity fields, and the mean value of positioning error is 620.72 m.