Research On Ship Instantaneous Linear Motion Based SINS Navigation Solution And Combination Forecasting

Ship swaying in six degrees excited by marine environment, such as winds and waves, will significantly decrease the accuracy of on-board weapon or other devices, and bring negative effects on the ship equipment and members. In order to avoid the accidents during taking off and landing of the aircraft, fighting path of the aircraft should be adjusted according to the motions of carrier, but real-time oscillations along three axes should be extracted and predicted over a period firstly. In this paper, the methods about extraction and prediction of the ship swaying motions are investigated.Firstly, the ship movements sailing on the sea are studied in detail, and the generating mechanism and modelling methods about ocean waves are discussed and a specific model of ship swaying motion is selected which is more closer to engineering than those methods based on a single frequency. The navigation principles of the Strapdown Inertial Navigation System (SINS) are described and the improved method to measuring oscillations based on SINS solution is detailedly deduced; an extraction method based on digital filter is proposed based on the analysis on the frequency components of solution results of SINS.Secondly, the characters about IIR and FIR digital filters are studied and compared, and IIR is selected as the tool to extracting the oscillation motion for the merits of shorter delay time than that of FIR. Working mechanism and design process are introduced and the parameter selection principles based on the analysis of frequency component are detailedly described. Simulations are executed to verify the effectiveness of the proposed filter.Finally, the methods of neural network (NN), for the merits of nonlinear, adaptive learning ability and data fusion, are selected to predict the ship oscillation for characters of stochastic and nonlinearity of ship oscillation. Three kinds of NNs, namely BP, RBF and combination of BP-RBF, are analyzed and compared with simulations to cope with the prediction of ship oscillation. Simulation results indicated that the combination method owns the merit of fast learning ability inherited from RBP and the merit of generalization ability from BP. Thus, it can be concluded that combination method can improve the prediction speed and ensure the predicting accuracy, which excels the uncombined methods in some extent.