Research On Integrated Navigation Technology Of Underwater Vehicle Based On ANFIS

Nowadays, as the most important exploration tool for developing and using ofmarine space, underwater vehicles get more and more applications. The navigationsystems and related technologies that underwater vehicles use become aninternational research hotspot. The navigation systems on the underwater vehiclealmost are strap-down inertial navigation system(INS), which should work underthe requirements of long-endurance and high-precision, this paper introducesadaptive network fuzzy inference system(ANFIS) to study the methodsof integrated navigation based on inertial navigation system and other navigationsystems, as well as the navigation error suppression technology after theintroduction of the damping network to the inertial navigation system, the purposeis to enhance the actual performance of the strap-down inertial navigation system.Our major works as follows:Firstly, adaptive network fuzzy inference system (ANFIS) is introduced basedon the introduction of artificial neural networks and fuzzy inference system, whichhas the advantages of the above two methods. The focuses of our analysis are thestructure mechanism and characteristics of the ANFIS, based on its characteristicswe introduce the areas that ANFIS can be used. And this paper study methods ofimproving navigation accuracy of inertial navigation system based on adaptivenetwork fuzzy inference system.Secondly, we introduce the principle of inertial navigation system, GPSsatellite navigation system, and summarize their characteristics. For the integratednavigation system which combines the advantages of both, the defect of thetraditional method using the Kalman Filter is obvious. So a proposed method ofINS/GPS data fusion when GPS signal is interrupted is adopted in this paper,ANFIS trains data when GPS information can be used, the correction informationthat is provided by ANFIS is given to inertial navigation system when the GPSinformation can not be used, and the effect of this method is verified by simulation.The analysis of Doppler Velocity Log (DVL) and Electromagnetic Log (EML)and hoe they works are done, according to the characteristics of two kinds of speedinformation that are provided by DVL and EML respectively, a proposed approachthat using the speed difference between of Doppler Speed Log and Electromagnetic Log estimate the ocean current speed of the current waters. Speed information ofDoppler Speed Log and Electromagnetic Log can be used when the data trainingusing ANFIS, and output of ANFIS is the estimated speed error of ocean currentspeed, estimated effect of the method is verified by simulation.Finally, the characteristics analysis of the different working states of theinertial navigation system is done, proving the effectiveness of introducing dampingnetwork in pure inertial navigation system. A new method is proposed for dampinginertial navigation system, which is, introducing ANFIS to intelligently transformdamping coefficient of damping inertial navigation. A different state of motioncorresponds to an optimal damping coefficient respectively. The simulation of thecorrespondence between the set of the damping coefficient and underwater vehiclemotion state should be done by using MATLAB, then ANFIS could use the data forits training, then ANFIS could provide the best damping coefficient corrodingunderwater vehicle motion, so that the damping inertial navigation system couldwork in the best state, navigation accuracy of damping inertial navigation systemshould be improved. The method is validated by simulation using inertial navigationsystem precision.