Research On Inertial Measurement Method Of Hull Deformation And Noise Suppression With Kalman Filter

In order to achieve the coordination among all the observation and combat equipments of modern ships,it is necessary to establish a unified spatial coordinate system,or namely the unified attitude reference system of the whole ship.As the hull deformation is an important factor affecting the establishment of a unified reference for each battle position point,its accurate measurement is an indispensable means to achieve high-precision attitude reference.At present,there are many measurement technologies for hull deformation angle.Inertial matching method has become a research focus because of its advantages such as real-time,low cost and high precision,et.al..Its core idea is to capture outputs of two sets of fiber gyroscope unit(FGU)through a synchronous recording device,and then get the deformation estimation results by Kalman filter of a host computer software.Based on the inertial matching technology of hull deformation,this paper lucubrates more inertial matching methods and several high-precision Kalman filters in the environment interfered by noise model mismatch,unknown noise,and non-Gaussian noise,which are proposed to improve performance of the measurement system.This paper contains several following key points:(1)Starting from the basic solving methods of deformation,four inertial matching methods are introduced with their algorithm formulas in detail,and the corresponding iterative equations of Kalman filter are established.According to common indicators,several groups of results of experiments through distinct swaying excitations and hull deformations have verified the effectiveness of the four algorithms.Besides,the weak observation verification experiment also tests their estimation repeatability of deformation.Eventually,their advantages or disadvantages will be discussed,which provide a variety of alternative schemes for hull deformation measurement.(2)Aiming at the performance deterioration of deformation estimation caused by unknown noise of filtering system and model mismatch of noise covariance matrix,this paper proposes to absorb the idea of adaptation into the conventional Kalman filter and design an adaptive deformation measurement system as well as providing the detailed algorithms’reproduction process through multi-slice improvement of adaptive iteration process.The experimental results also show that the improved algorithm plays effective roles in anti-interference performance and adaptive performance,and can achieve better deformation measurement accuracy under complex external disturbances.(3)Aiming at the problem of big estimation deviation of conventional Kalman filter when the measurement system locates in non-Gaussian environment,the concept of Maximum Correntropy Criterion(MCC)as well as the principle of Weighted Least Squares(WLS)is embedded into the conventional Kalman filter,which gets improved and capture the ability of non-Gaussian filtering and robutness.The results of several groups of horizontal and vertical comparative experiments have not only confirmed that the maximum correntropy filter can achieve the optimal filtering but satisfactory performances of accuracy and convergence,as well as stronger robustness and practicability,which provides specific guidance for the selection of adaptive and maximum correntropy in different circumstances.