Research Of Underwater Vehicle Gravity Matching Algorithm

Gravity matching algorithm is the kernel of gravity aided inertial navigation system, it makes correlation analysis to gravity anomaly data and the output of inertial navigation information according to certain method. And then the estimation about the carrier position will be obtained to correct the error of inertial navigation. The real-time performance of traditional sequence matching algorithms is bad and the single point matching algorithm is easily to divergence in the matching area with large gravity anomaly changes. And the adjacent matching points of traditional matching algorithm are independent of each other, therefore the mismatching results can’t be revised in time. Thus improve the matching accuracy and ensure the credibility of the matching results for long-endurance inertial navigation system error correction has important significance. In this paper, the vector matching algorithm which adds the position correlation between adjacent INS sample points into the matching based on particle filter. The proposed method overcomes the mismatching shortcomings of the traditional methods with higher accuracy and robustness. The main work and innovation points are as follows:(1) The particle filter takes place of the extended Kalman filter is put forward. The traditional single point matching algorithm is easily divergence in the matching area with large gravity anomaly changes due to the linearization error. Aiming at this problem, using particle filter instead of extended Kalman filter to avoid linearization error and expand the scope of matching algorithm application. The simulation experiments show that this method can avoid the disadvantage of SITAN matching algorithm in the matching area with large gravity anomaly changes and improve the matching accuracy.(2) The fixed scales and orientations vector matching algorithm is proposed based on the particle filter single point matching algorithm. The proposed method can avoid mismatching of traditional matching algorithm. The distant correlation of adjacent INS sampling points is added into the single point particle filter results. The current matching result is revised by the selected surrounding sampling point results. The final matching result is obtained by weighted least squares to reduce the random error and using the greedy algorithm for reference. The simulation experiments show that the proposed algorithm has higher precision and stronger robustness compared with the traditional matching algorithms.(3) To extend the scope of algorithm application, the mutative scales and orientations vector matching algorithm is proposed based on fixed scales and orientations vector matching algorithm. In order to adapt the maneuverability of underwater vehicle, the credibility discriminant is improved to suit the not uniform and nonlinear sailing. And to improve the accuracy of algorithm in the bending of vehicle, the greater moment of sampling points are added into rectifying the current moment matching point. And the phase correlation between adjacent INS sampling points is added into the matching process. The simulation results show that the proposed algorithm has more accuracy and robustness for maneuvering navigation.