Research On Key Techologies Of INS/TAN/GAN Integrated Navigation System For Underwater Vehicles

Navigation technology is one of the key technologies of underwater vehicle in the process of design and development.As the marine environment is complex and changeable,to make the vehicle can be independent,efficient,safe and self-contained navigation tasks are still confronted with many challenges.At present,the underwater vehicle mostly rely on inertial navigation system(INS)mainly,supplemented by GPS,acoustic equipment for realizing navigation task,but found that this kind of navigation mode has many deficiencies in the long period of application.Therefore,geophysical navigation technology becomes an important research direction,which includes terrain aided navigation(TAN),geomagnetic aided navigation(GAN)and gravity aided inertial navigation(GAIN).Although the INS/TAN,INS/GAN,INS/GAIN research has been more common,the use of mixed-assistant positioning research seldom reported.This paper takes the underwater vehicle as the research object,combined with the feature of terrain and geomagnetic,then focus on the key technologies of INS/TAN/GAN system design.In fact,the integrated navigation positioning accuracy is closely related to the suitability of the passing area of underwater vehicles.The suitability of any area can be expressed by terrain or geomagnetic standard deviation,the characteristic parameters of roughness and correlation coefficient,and then a combined parameter is put forward by the ratio of standard deviation and range for describing the relative fluctuation massif.Based on fuzzy theory,terrain and geomagnetic characteristic parametersare selected as the evaluation index,and are calculated the weight of each index by the method of CRITIC firstly,and then it is established the terrain/geomagnetic data suitability analysis model.Finally,two methods including fusion first and then evaluation,or first evaluation and then fusion are proposed in the thesis.It also puts forward a novel projecting pursuit-based selection method with the terrain/geomagnetic data suitability analysis,and respectively by adopting the particle swarm algorithm,differential evolution algorithm and the firefly algorithm for the objective function optimization.After the selection of all characteristic parameters as indicators,it discusses the cases of the best projection direction distribution.Finally,it gives the experimental verification with the same data in the second chapter.According to the characteristics of underwater vehicle working environment,an INS/TAN/GAN integrated navigation scheme is proposed.First of all,combined with the principle of INS,TAN and GAN,it adopts the error propagation model of INS as the system state,with navigation results of TAN and GAN as observation,then established mixed-assistant positioning model.Compared to the pure INS/TAN,INS/GAN,the algorithm has more advantages in the stability,and most the estimation results are more reliable.But on the conditions that the terrain or geomagnetic matching error is large,the mixed filtering results also become worse.In order to make the INS/TAN/GAN navigation system achieve the best performance,a federal integrated navigation scheme is proposed in the chapter four.Using a reset federal filter structure,the INS/TAN,INS/GAN are set as the navigation subsystem.Simulation experiments are executed with the standard federal filter and the adaptive Sage-Husa Kalman filter respectively,and the results show that the adaptive method can be better for the environment.Then,a method is put forward to extract reference prior navigation information based on TAN/GAN map,and the validity of the observation noise prediction method with the TAN/GAN suitability is proven at last.Sage-Husa adaptive filter is to adjust the system observation noise by using the change of prediction residuals,and when the change is not entirely caused by the prediction residual measurements,the validity of the method is greatly limited.Aiming at this problem,it puts forward the Sage-Husa adaptive filtering algorithm based on the adaptive suitability optimization with the terrain/geomagnetic reference map,which is by using the suitability database of reference map and the statistics relationship results of the suitability with the matching error,combined with the carrier track position sequence,and then predict the next position observation noise.This method can estimate the matching error directly by avoiding misalignment estimation problem in traditional Sage-Husa adaptive filtering.Aiming at the problem of fault detection of the integrated navigation system,it designs a federal combined fault tolerant structure to improve the flexibility and reliability of integrated navigation system,and the performance of the structure is verified by simulation experiments finally.