Research Of Algorithms On Gravity Matching Aided Inertial Navigation System

As a country which has so many islands and a large area of ocean, China puts more emphasis on the development of ocean technology because the role of ocean development strategy is more and more important. As an indispensable step of developing ocean technology, the development of underwater vehicles(such as the Jiaolong) has a strategic significance. Inertial navigation system(INS) is widely used in the underwater vehicle’s navigation system. INS has the advantages of passivity and autonomy which offer the conditions for concealment. However, INS has the problem of error accumulation.Therefore, INS cannot ensure the high precision of navigation for the long-endurance underwater vehicles. It is necessary to correct the error of INS with the assistance of other navigation methods. Gravity matching aided inertial navigation system utilizes the gravity information which is the inherent information of the earth. It can correct the drift of INS in the condition of long-endurance, which satisfy the requirement of the underwater vehicles.In this paper, gravity matching aided inertial navigation system is considered as the research object and researches of algorithms on gravity matching and Kalman filter have been done. The main work is:(1) The real-time performance of the traditional algorithm of correlative extreme value is poor. Aiming at this problem, this paper proposed an improved algorithm. The improved algorithm improves the traditional algorithm from two aspects which are the introduction of variable searching area and the removal of interference data. The improved algorithm uses the variable searching area instead of the fixed searching area. The searching area changes with the variation of the INS position’s drift distance, which decreases the searching range.Then the improved algorithm finds the points whose gravity anomaly value is close to the data from the gravimeter, which accomplishes the extraction of effective alternative matching points. What’s more, because the distance between the matching trajectory point and the INS trajectory point is bigger and bigger, the constraint condition is introduced. The constraint condition decreases the alternative matching trajectories. In consideration of the drift of INS gradually increases, the improved algorithm introduces the weighted value. It means that the data with higher accuracy has a lager weight. The results of simulations show that the improved algorithm has the excellent matching precision and improves the matching efficiency.(2) Aiming at the problem of unequal-interval when the matching trajectory and the INS trajectory are combined, this paper designs the unequal-interval filter algorithm whose measurements are the differences between the matching trajectory points and the INS trajectory points. At first, this paper makes the credible judgment of the matching trajectory using the gravity field characteristic parameters, and then gets the unequal-interval characteristics of the matching trajectory. When gravity characteristics of matching area vary clearly(the continuous missing distance is less than 15 nautical miles), the process of Kalman filter only has time update when the matching trajectory point is missing. When there is an output of the matching trajectory point, the filter process has both time update and the measurement update. When gravity characteristics of matching area vary unclearly(the continuous missing distance is 15 nautical miles or more than 15 nautical miles), this paper uses cubic spline interpolation to insert the estimates in the positions of the missing matching trajectory points and then do the fusion. Besides, this paper uses the method of extending measurements to deal with the problem of colored measurement noise. The results of simulations show that the unequal-interval filter algorithm can solve the unequal-interval problem and improve the navigation accuracy.(3) This paper establishes the semi-material simulation platform based on gravity matching aided inertial navigation system by MATLAB/GUIDE. This platform contains four modules which are "Data Importing", "Route Planning", "Gravity Matching Algorithm" and "Filter Fusion Algorithm". This platform has a complete architecture and a clear interface. What’s more, the platform can be transplanted and expanded easily and the platform lays the foundation of intensive research of gravity matching aided inertial navigation system.