Research On The Method Of Underwater Gravity-aided Navigation And Positioning

Gravity aided navigation is an effective method for submarine and other underwater vehicle to correct the accumulation error of inertial navigation systems(Inertial Navigation System,INS).Unlike satellite/radio navigation and terrain-aided navigation,gravity aided navigation relaying on the passivity of gravity measurement,can ensure the reliability of navigation and the concealment of carrier,while limiting the growth of INS error.Therefore,gravity aided navigation has very high value in military.Gravity aided navigation is four modules compose system:INS,gravity measurement instrument,gravity reference map and matching location algorithm,the performance and interaction between modules will affect the accuracy of gravity matching.This paper studies the modules of gravity-aided navigation,according to the characteristics of underwater vehicle and the Earth's gravitational field.This paper explains the benchmark role of INS in gravity-aided navigation system;analyses the error factors of gravity measuring instrument and gives the treatment method of gravity measurement error.Studies the interpolation algorithm of gravity reference map,Kriging advanced Shepard,the radial basis function and green function spline algorithm.Points out that,with appropriate smoothing factor,the interpolation accuracy of radial basis function is higher.In addition,aiming at the problem of large interpolation error of high frequence area in reference map,the paper proposes to identify the key measuring area(suspected high-frequency area)of reference map by wavelet transform,and focus on the key measuring area to increase measuringpoints with random sampling strategy in order to reduce the interpolation error.Theinterpolation experiment result shows that,this method can effectively improve interpolationaccuracy of reference map,and requires less measuring points increased,has less extreme valueof interpolation error,so has higher efficiency.Gravity aided navigation is an auxiliary navigation method based on INS,the positioningaccuracy of INS directly affects the positioning effect of matching algorithm.This paperanalyzes the correlation between several kinds of classic correlation and recursive filteringgravity-matching algorithms with the growth rate of INS position error from the perspective ofalgorithm principle.The result of the analysis points out that the improvement of positionaccuracy of INS has different extent of positioning accuracy improvement for different matching algorithm,and could change the excellent order between algorithms.Then,some matching experiments were done with 3 kinds of INS,whose accuracy changed gradually from moderate accuracy to high accuracy,to verifies the correctness of analysis,and draws the criterion for selecting matching algorithms according to the accuracy of INS.The strategy for precision improvement of the maximum correlation-matching algorithm and the recursive filtering matching algorithm is proposed.This strategy evaluates the matching effect,and decides to reject and select the matching information by the standard deviation of correlation value,the error lower bound of filtering matching and customizing criterion.The simulation result verifies the effectiveness of the strategy.Based on the definition of the parameters of absolute gravity field's local characteristics,the paper proposes to select the initially navigation area by a certain threshold.Then,it extracts the simplified skeleton points of matching areas by fast Euclidean distance field algorithm and skeleton point's simplification algorithm,and divides the original navigation area into different local area by the simplified skeleton points and their Euclidean distance value.By calculating the gravity field statistical characteristics of local areas,the navigability and the heading suitable for gravity-matching is confirmed in order to provide the basis for selecting the carrier heading.Gravity matching experiment results in different local areas show the effectiveness of the local analysis methods.In addition,the empirical criterion is drawn from the experiment results to judge the carrier's heading,which is conducive to matching,according to gravity field slope roughness and standard deviation.Taking gravity gradient as the background,this paper proposes a correlation-matching algorithm,which is applicability for matching in the large position error condition.Based on the segmented short tracks of INS,the algorithm gets the matching track with the path penalty function and shortest path methodm,inorder to reduce the matching error caused by the INS accumulated error.In addition,this paper uses STUPF algorithm for small position error filter matching to deal with the nonlinear characteristics of gravity gradient and the impact of model error of filtering system to position accuracy.Then,this paper divides the filtering system into linear and nonlinear parts,and introduces STU_RBPF linear/nonlinear mixed filter algorithm.The simulation result verifies the superiority of the algorithm relative to STUPF,and it confirms that,even for moderate accuracy INS,the gravity gradient matching has higher positioning accuracy than the absolute gravity field matching of high accuracy INS,which also verifies the advantage of gravity gradient relative to absolute gravity in matching and positioning.