Research On Submarine Terrain Aided Navigation Technology Based On Particle Filter

Autonomous Underwater Vehicle(AUV)has become an important tool for exploring and developing the ocean,and high-precision underwater navigation is an important guarantee for AUV’s underwater operations.Therefore,high-precision and highly autonomous underwater navigation Technology has become a difficult and hot issue in the field of navigation technology.Seabed Terrain Aided Navigation(STAN)uses the elevation information of the seabed terrain to achieve navigation and positioning.It does not require additional sensors in the system and does not require the submersible to float regularly to correct errors.It can meet the requirements of AUV high precision,high concealment,high autonomy and long endurance navigation.Therefore,this article takes the high-precision navigation and positioning of AUV during navigation as the research background to study the submarine terrain assisted navigation and positioning technology.This paper first studied the problem of submarine terrain modeling,using bilinear interpolation algorithm to model the submarine terrain,and obtain a digital elevation map with available resolution.Introduce the basic navigation system and the working principle of Doppler log.Perform error analysis,theoretical derivation and error model establishment for inertial navigation system and Doppler log.Secondly,analyze the main errors of the submarine terrain assisted navigation system,analyze the sounding data errors of the multi-beam sounding system,and filter the measurement data.From the three aspects of algorithm principle,matching algorithm core and system characteristics,the terrain contour matching system(TERCOM),Sandia system(SITAN)and ICP algorithm are introduced,and the advantages and disadvantages of the three algorithms are analyzed.Then,in order to meet the accuracy requirements of AUV underwater navigation,an underwater SINS/DVL integrated navigation system was designed,and the stability of the system was verified by simulation.On the basis of the SINS/DVL integrated navigation system,the TAN/SINS/DVL loose integrated navigation system was designed,and the application of particle filter in the loose integrated navigation system was studied,and simulation verification was carried out respectively.In the application of nonlinear measurement models,especially at the inflection point where the sign of the slope changes,measurement uncertainty is the key reason for filter degradation and divergence.Aiming at the problem of local measurement ambiguity in particle filtering,an improved particle filter is proposed and simulation analysis is carried out.Finally,in order to overcome the various shortcomings of the traditional single evaluation index in the evaluation of the seabed terrain,a comprehensive evaluation method based on fuzzy decision theory is proposed,which integrates terrain roughness,terrain height standard deviation,terrain correlation coefficient and terrain A variety of parameter evaluation indicators such as entropy fully reflect the characteristics of the topography,the amount of information,and the roughness in a specific area.The terrain matching algorithm uses a particle filter algorithm,and the simulation results verify the effectiveness of this method for adaptability analysis of seabed terrain.