A Gravity Matching Adaptability Analysis Method Based on Gravity Directional Characteristics

Gravity aided inertial navigation system is one of the key technologies for underwater navigation and positioning. The position accuracy relates to the spatial variability and directional characteristics of the gravity anomaly map. To quantify the directional adaptability of gravity anomaly maps, this paper proposes a novel analysis method based on gravity directional characteristics. Firstly, three key parameters are designed: gravity profile similarity, directional gravity gradient, and directional gradient concentration, which characterize the directional distribution of gravity anomalies. Then, a balanced gravity characteristic parameter is constructed by fusing classical gravity metrics. The mapping between these features and directional matching rates is then modeled using an LM-BP neural network, establishing an adaptability evaluation framework. Experimental results demonstrate that in regions with directional adaptability exceeding 0.8, the matching rates are over 80%. The proposed model offers higher accuracy and comprehensiveness in selecting gravity matching areas than existing methods and provides a practical tool for optimizing gravity matching accuracy.