Research On Adaptive Filtering Of INS/DVL Integrated Navigation And In-motion Fast Alignment For Underwater Vehicle

The integrated navigation system which is composed of DVL, INS and Kalman filter is an ideal underwater navigation system. The paper mainly deals with adaptive filter of INS/DVL integrated navigation and in-motion fast alignment for underwater vehicle based on DVL’s work and error characters.The main contents and work include the following aspects:(1)An fuzzy adaptive filter considering bias jump of DVL measurement error is proposed on this paper. Compared with the present fuzzy adaptive filter algorithm, the proposed algorithm introduces bias jump adaptive filter to compensate the bias of DVL measurement error in real-time. By monitoring the changing of the filter residual average and the difference between the filter residual covariance with the theoretical filter residual covariance in real-time, membership function can be adapted to modify the statistical properties of the measurement noise and DVL measurement singularity can be eliminated. So the precise and adjustability of INS/DVL integrated navigation algorithm can be improved. This paper uses the difference between the filter residual covariance with the theoretical filter residual covariance as fuzzy input to overcome conventional fuzzy adaptive algorithm’s defect of being much too sensitive to the change of residual covariance.(2)The paper analyzes three different backtracking alignment schemes and their characters, implement processes in inertial frame-0bi, earth fixed frame-0n, navigation frame- n. As is shown through vessel aligning results, three backtracking alignment schemes can reach the same alignment precise and are the same in essence, but they have different implement complexity. Considering the program of algorithms, backtracking alignment scheme in earth fixed frame-0n is more concise.(3)Contrapose the big DVL measurement error and measurement singularity, put out an in-motion fast alignment for underwater vehicles using fuzzy adaptive filter based on backtracking alignment scheme. By using measurement information in real-time, the statistical properties of the measurement noise can be modified and DVL measurement singularity can be eliminated. As is shown through vessel aligning results, fuzzy adaptive filter algorithm can guarantee the alignment precise, even though DVL measurement error is big and DVL exists measurement singularity. While there has been a sharp decline in alignment precise of Kalman filter algorithm.(4)The thesis constructs INS/DVL integrated navigation system and achieves real time multi-tasking integrated navigation system software based on Vx Works 5.5 for Power PC platform. Whole design of navigation system is conceived, software’s work pattern and work procedure are analyzed. Due to the navigation software’s functions, multi-tasking modules are be delimited and two different synchronous schemes are conceived. An offline test proves INS/DVL integrated navigation algorithm’s practicability.